feat(radar): stable catalog export workflow with provenance (#10826 )

Passo 10 of the Radar go-live: publish the OmniRoute catalog export the Radar server consumes to a stable URL, so the 1 GB-RAM private server pulls it via RADAR_EXPORT_URL instead of depending on the deploy-time snapshot. - scripts/release/radar-export.mjs: emits {geradoEm, budgets, totais, registry, provenance} from the catalog config modules. Provenance (sourceCommit, sourceRef, runUrl, generatedBy) is never fabricated — unknown fields stay null. - .github/workflows/radar-export.yml: on main catalog changes / manual dispatch / weekly, generates the export and clobbers the stable 'radar-export-latest' release asset (gh release, GH_TOKEN — checkout persist-credentials:false). - tests/unit/radar-export.test.mjs: consumer contract (budgets[] non-empty) + provenance null-when-unknown + GitHub-env reflection. Stable URL for RADAR_EXPORT_URL: https://github.com/diegosouzapw/OmniRoute/releases/download/radar-export-latest/export-omniroute.json Re-baselines zizmorFindings 190->192 (+2 unpinned-uses @vN, the repo-wide deliberate convention; artipacked auto-fixed). Co-authored-by: Xiangzhe <bakryun0718@proton.me>
fix(resilience): scope same-account transport retry out of emergency-fallback and combo hops
2026-08-20 22:22:57 +03:00 · 2026-08-20 08:09:42 -03:00 · 2026-08-20 08:08:09 -03:00 · 2026-08-20 06:37:02 -03:00 · 2026-08-20 06:31:49 -03:00 · 2026-08-20 06:30:46 -03:00
226 changed files with 9782 additions and 1502 deletions
--- a/.env.example
+++ b/.env.example
@@ -1033,6 +1033,10 @@ PROVIDER_LIMITS_SYNC_SPACING_MS=1500
 # Used by: src/lib/db/core.ts::getDbHealthCheckIntervalMs().
 #OMNIROUTE_DB_HEALTHCHECK_INTERVAL_MS=21600000

+# WAL truncate cadence override (ms).  Set to 0 to disable.  Default: 21600000 (6h).
+# Used by: src/lib/db/core.ts::getWalTruncateIntervalMs().
+#OMNIROUTE_WAL_TRUNCATE_INTERVAL_MS=21600000
+
 # Skip the Redis-backed auth cache used by API key lookups (forces DB reads).
 # Used by: src/lib/db/apiKeys.ts.  Set to 1 to disable.  Default: enabled.
 #OMNIROUTE_DISABLE_REDIS_AUTH_CACHE=0
--- a/.github/workflows/radar-export.yml
+++ b/.github/workflows/radar-export.yml
@@ -0,0 +1,64 @@
+# Publica o export estável do catálogo consumido pelo OmniRoute Radar numa URL
+# fixa (asset de release `radar-export-latest`), para o servidor privado do Radar
+# (1 GB RAM, nunca clona/builda o OmniRoute) baixá-lo via `RADAR_EXPORT_URL` em
+# vez de depender do snapshot gravado no deploy. Fonte: scripts/release/radar-export.mjs.
+#
+# A URL estável resultante (definir em RADAR_EXPORT_URL no .env do radar-server):
+#   https://github.com/diegosouzapw/OmniRoute/releases/download/radar-export-latest/export-omniroute.json
+name: Radar Export
+
+on:
+  workflow_dispatch: # o operador pode publicar sob demanda (de qualquer ref)
+  push:
+    branches: [main] # produção: só o catálogo do main clobra o asset estável
+    paths:
+      - open-sse/config/freeModelCatalog.data.ts
+      - open-sse/config/freeModelCatalog.ts
+      - open-sse/config/providerRegistry.ts
+      - open-sse/config/providers/**
+      - scripts/release/radar-export.mjs
+      - .github/workflows/radar-export.yml
+  schedule:
+    - cron: "17 6 * * 1" # semanal (segunda 06:17 UTC): mantém geradoEm/proveniência frescos
+
+permissions:
+  contents: read
+
+concurrency:
+  group: radar-export-${{ github.ref }}
+  cancel-in-progress: true
+
+env:
+  CI_NODE_VERSION: "24"
+
+jobs:
+  publish-export:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write # gh release upload — clobra o asset estável do export
+    steps:
+      - uses: actions/checkout@v7
+        with:
+          persist-credentials: false # publish usa GH_TOKEN via gh release, não a credencial do checkout
+      - uses: actions/setup-node@v7
+        with:
+          node-version: ${{ env.CI_NODE_VERSION }}
+          cache: npm
+      - run: npm ci
+      - name: Generate catalog export with provenance
+        run: node --import tsx/esm scripts/release/radar-export.mjs "$RUNNER_TEMP/export-omniroute.json"
+      - name: Publish to the stable release asset
+        env:
+          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          set -euo pipefail
+          TAG="radar-export-latest"
+          # Cria o release estável na primeira vez; nas seguintes só re-anexa o asset.
+          if ! gh release view "$TAG" --repo "$GITHUB_REPOSITORY" >/dev/null 2>&1; then
+            gh release create "$TAG" \
+              --repo "$GITHUB_REPOSITORY" \
+              --title "Radar catalog export (rolling)" \
+              --notes "Export estável do catálogo OmniRoute para o Radar. Atualizado automaticamente; NÃO é um release de versão do produto." \
+              --latest=false
+          fi
+          gh release upload "$TAG" "$RUNNER_TEMP/export-omniroute.json" --repo "$GITHUB_REPOSITORY" --clobber
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -46,7 +46,7 @@ Repository map and Reference Documentation sections below.

 ## Project at a Glance

-**OmniRoute** — unified AI proxy/router. One endpoint, 341 LLM providers, auto-fallback.
+**OmniRoute** — unified AI proxy/router. One endpoint, 342 LLM providers, auto-fallback.

 | Layer         | Location                | Purpose                                                                                                                                                                   |
 | ------------- | ----------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
--- a/README.md
+++ b/README.md
@@ -7,7 +7,7 @@

 # 🚀 OmniRoute — The Free AI Gateway

-<img src="./docs/diagrams/readme-hero.svg" width="100%" alt="OmniRoute — Never stop coding. Every AI tool → 341 providers — 90+ free — through one endpoint. Claude Code, Codex, Cursor, Cline, Copilot & Antigravity into FREE Claude / GPT / Gemini with auto-fallback. RTK + Caveman stacked compression saves 15–95% tokens (~89% avg) — never hit limits. 341 AI providers · 90+ free tiers · ~1.51B free tokens/mo · 19 routing strategies · $0 to start."/>
+<img src="./docs/diagrams/readme-hero.svg" width="100%" alt="OmniRoute — Never stop coding. Every AI tool → 342 providers — 90+ free — through one endpoint. Claude Code, Codex, Cursor, Cline, Copilot & Antigravity into FREE Claude / GPT / Gemini with auto-fallback. RTK + Caveman stacked compression saves 15–95% tokens (~89% avg) — never hit limits. 342 AI providers · 90+ free tiers · ~1.51B free tokens/mo · 19 routing strategies · $0 to start."/>

 </div>

@@ -63,7 +63,7 @@

 |                           | v3.8.49 | **v3.8.50** | `v3.8.51+`  |
 | ------------------------- | :-----: | :---------: | :---------: |
-| 🌐 Providers              |   290   |   **341**   | more queued |
+| 🌐 Providers              |   290   |   **342**   | more queued |
 | 🧠 Documented models      |  1185   |  **1202**   |      —      |
 | 🖼️ Modality Bridge        |    —    |  🆕 vision  |    video    |
 | 📡 Radar free catalog     |    —    |  🆕 opt-in  |      —      |
@@ -101,7 +101,7 @@
  <tr>
    <td align="right"><b>⚙️ Features</b></td>
    <td align="center"><a href="#-combos--the-flagship">🎯 Combos</a></td>
-    <td align="center"><a href="#-341-ai-providers--90-free">🌐 Providers</a></td>
+    <td align="center"><a href="#-342-ai-providers--90-free">🌐 Providers</a></td>
    <td align="center"><a href="#-full-cli--a2a--mcp">🔌 CLI &amp; MCP</a></td>
  </tr>
  <tr>
@@ -210,7 +210,7 @@ curl http://localhost:20128/v1/chat/completions \

 </div>

-<img src="./docs/diagrams/promise-pillars.svg" width="100%" alt="The Promise — One endpoint. 341 providers. Never stop building — OmniRoute picks the cheapest one that works. Six pillars: Never hit limits (auto-fallback across 341 providers in milliseconds, zero downtime) · Save up to 95% tokens (RTK + Caveman stacked compression cuts 15–95%, ~89% avg on tool-heavy sessions) · $0 to start (90+ free tiers, 56 free forever — no card needed) · Every tool works (33 coding agents through one config) · One endpoint (OpenAI ↔ Claude ↔ Gemini ↔ Responses API at /v1) · Production-grade (circuit breakers, TLS stealth, MCP 109 tools, A2A, memory, guardrails, evals — 25,000+ tests)."/>
+<img src="./docs/diagrams/promise-pillars.svg" width="100%" alt="The Promise — One endpoint. 342 providers. Never stop building — OmniRoute picks the cheapest one that works. Six pillars: Never hit limits (auto-fallback across 342 providers in milliseconds, zero downtime) · Save up to 95% tokens (RTK + Caveman stacked compression cuts 15–95%, ~89% avg on tool-heavy sessions) · $0 to start (90+ free tiers, 56 free forever — no card needed) · Every tool works (33 coding agents through one config) · One endpoint (OpenAI ↔ Claude ↔ Gemini ↔ Responses API at /v1) · Production-grade (circuit breakers, TLS stealth, MCP 109 tools, A2A, memory, guardrails, evals — 25,000+ tests)."/>

 <br/>
 <br/>
@@ -461,7 +461,7 @@ All **19** strategies — mix & match per combo step:

 </div>

-<img src="./docs/diagrams/comparison-table.svg" width="100%" alt="What sets OmniRoute apart — comparison table vs 9router, OpenRouter, CLIProxyAPI and LiteLLM across 13 capabilities. OmniRoute: 341 providers, 90+ free providers built-in, 19 routing strategies, 12-engine token compression, built-in MCP server with 109 tools, A2A agent protocol, persistent memory, guardrails, cloud agents, TLS fingerprint stealth, Desktop/Termux/PWA, 43 i18n UI locales, 100% MIT self-hosted. OmniRoute is the only one with the full set; competitors show a mix of checks, partials and crosses. Verified from each project&apos;s docs."/>
+<img src="./docs/diagrams/comparison-table.svg" width="100%" alt="What sets OmniRoute apart — comparison table vs 9router, OpenRouter, CLIProxyAPI and LiteLLM across 13 capabilities. OmniRoute: 342 providers, 90+ free providers built-in, 19 routing strategies, 12-engine token compression, built-in MCP server with 109 tools, A2A agent protocol, persistent memory, guardrails, cloud agents, TLS fingerprint stealth, Desktop/Termux/PWA, 43 i18n UI locales, 100% MIT self-hosted. OmniRoute is the only one with the full set; competitors show a mix of checks, partials and crosses. Verified from each project&apos;s docs."/>

 <sub>📊 Full methodology &amp; per-feature detail vs 9router, OpenRouter, CLIProxyAPI &amp; LiteLLM → [`docs/comparison/OMNIROUTE_VS_ALTERNATIVES.md`](docs/comparison/OMNIROUTE_VS_ALTERNATIVES.md)</sub>

@@ -559,7 +559,7 @@ the current catalog at **[radar.omniroute.online/planos](https://radar.omniroute
 - **🖼️ New endpoints** — `/v1/ocr` (Mistral OCR) and `/v1/audio/translations` (Whisper-style) round out the media surface. → [API Reference](docs/reference/API_REFERENCE.md)
 - **🎨 Image / video / audio generation** — one API for media: xAI Grok Imagine & Novita AI video, ComfyUI, Freepik, Adobe Firefly, Microsoft Designer, Segmind, EdgeTTS. → [API Reference](docs/reference/API_REFERENCE.md)
 - **🌍 Deployment & ops** — reverse-proxy `basePath`, browser-language auto-detect, per-key device tracking, root-less MITM trust, zh-TW localization. → [Environment](docs/reference/ENVIRONMENT.md)
- **🤝 More providers & agents** — Cursor Cloud Agent, Grok Build (xAI) with browser + OAuth login, Ollama first-class card, Claude Opus 5 & Sonnet 5, Kimi official partnership (Code/Web/Moonshot), Zed, Requesty, SenseNova, Yuanbao, Agnes AI… and a refreshed **341-provider catalog**. → [Providers](docs/reference/PROVIDER_REFERENCE.md)
+- **🤝 More providers & agents** — Cursor Cloud Agent, Grok Build (xAI) with browser + OAuth login, Ollama first-class card, Claude Opus 5 & Sonnet 5, Kimi official partnership (Code/Web/Moonshot), Zed, Requesty, SenseNova, Yuanbao, Agnes AI… and a refreshed **342-provider catalog**. → [Providers](docs/reference/PROVIDER_REFERENCE.md)
 - **📡 Routing transparency** — every response carries an `X-OmniRoute-Decision` header naming the strategy/provider/latency that served it, a new `cache-optimized` combo strategy + Auto-Combo `cacheAffinity` factor route repeat requests back to the connection holding the cached prefix, and a read-only `/v1/auto-combo/{channel}/candidates` endpoint exposes an `auto/*` channel's live candidate pool. → [Auto-Combo](docs/routing/AUTO-COMBO.md)
 - **⚡ Local performance & infra** — one-click local Redis, Cloudflare Workers / Deno Deploy relay deployers, Bifrost & Mux as supervised embedded services. → [Embedded Services](docs/frameworks/EMBEDDED-SERVICES.md)

@@ -642,11 +642,11 @@ of your shell history. → [CLI Integrations](docs/guides/CLI-INTEGRATIONS.md)

 <div align="center">

-## 🌐 341 AI Providers — 90+ Free
+## 🌐 342 AI Providers — 90+ Free

 </div>

-> The most complete catalog of any open-source router: **341 providers**, **90+ with a free tier**, **56 free forever**.
+> The most complete catalog of any open-source router: **342 providers**, **90+ with a free tier**, **56 free forever**.

 <div align="center">

--- a/changelog.d/features/10729-cursor-api-key-and-cli-passthrough.md
+++ b/changelog.d/features/10729-cursor-api-key-and-cli-passthrough.md
@@ -0,0 +1 @@
+- **feat(providers):** new `cursor-api` provider (card "Cursor API", alias `cua`): connect a Cursor user API key (`crsr_…`) and route `cursor-api/<model>` through the existing Cursor agent executor (the key is exchanged for a 1h session token and cached), plus a `/api/cursor-cli/*` passthrough so the Cursor CLI itself runs through OmniRoute (`CURSOR_API_ENDPOINT=http://<omniroute>/api/cursor-cli`, `CURSOR_API_KEY=<OmniRoute key>`) with every RPC attributed and logged. The IDE `cursor` provider is unchanged. (#10729)
--- a/changelog.d/features/10771-health-root-endpoint.md
+++ b/changelog.d/features/10771-health-root-endpoint.md
@@ -0,0 +1 @@
+- **feat(api):** `GET /api/health` now answers `{ status, timestamp }` without a key. Until now the path had no route, so the management-auth boundary answered first with a 401 — indistinguishable from a wrong key or an unknown route, which left Docker HEALTHCHECKs and Kubernetes probes unable to tell "down" from "misconfigured". Kept deliberately minimal: version, uptime and memory stay behind the authenticated `/api/monitoring/health` ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10771)).
--- a/changelog.d/features/10783-task-routing-configurable-patterns.md
+++ b/changelog.d/features/10783-task-routing-configurable-patterns.md
@@ -0,0 +1 @@
+- feat(routing): make Task-Aware Smart Routing's detection patterns operator-configurable via `settings.taskRouting.patternOverrides` (`PUT /api/settings/task-routing`) — the built-in patterns are English-only, so a non-English dashboard had no recourse short of turning detection off entirely; an override now replaces the pattern list for one task type without touching the rest (#10783)
--- a/changelog.d/features/kimi-coding-extra-usage.md
+++ b/changelog.d/features/kimi-coding-extra-usage.md
@@ -0,0 +1 @@
+- **feat(usage):** show Kimi Coding's fixed-order Code 5-hour/7-day quota windows plus Extra Usage status, balance, monthly spend/limit, and the official Additional Credits link on Dashboard → Quota cards.
--- a/changelog.d/fixes/10732-copilot-m365-invocation-refresh.md
+++ b/changelog.d/fixes/10732-copilot-m365-invocation-refresh.md
@@ -0,0 +1 @@
+- **fix(providers):** copilot-m365-web chat turns no longer surface as `(empty response)` — the type:4 invocation is aligned with the 2026-08 wire shape and now carries its type:1 Metrics follow-up in the same socket write, and the access token pre-flight-refreshes from a stored refresh_token instead of requiring a DevTools re-capture every ~75 minutes ([#10732](https://github.com/diegosouzapw/OmniRoute/pull/10732) — thanks @acc0mplish)
--- a/changelog.d/fixes/10734-combo-context-generic-default.md
+++ b/changelog.d/fixes/10734-combo-context-generic-default.md
@@ -0,0 +1 @@
+- **fix(catalog):** stop counting `getTokenLimit()`'s generic 128k catch-all as a known combo window, so `/v1/models` advertises the min of sourced member contexts instead of collapsing a 500k combo to 128k ([#10734](https://github.com/diegosouzapw/OmniRoute/issues/10734))
--- a/changelog.d/fixes/10769-cache-stats-real-cache.md
+++ b/changelog.d/fixes/10769-cache-stats-real-cache.md
@@ -0,0 +1 @@
+- **fix(api):** `/api/cache/stats` reported the prompt-cache LRU, which no request path ever writes to — it answered `0 hit / 0 miss, size 0` while the semantic cache served real traffic, and the Health and Usage dashboards rendered that as fact. It now reports the semantic cache's in-memory entries, with the same response shape ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10769)) — thanks @Poid-ZA, who first fixed this in #9446.
--- a/changelog.d/fixes/10770-console-interceptor-message-fidelity.md
+++ b/changelog.d/fixes/10770-console-interceptor-message-fidelity.md
@@ -0,0 +1 @@
+- **fix(logging):** the app log is filterable and readable again. Entries from the tagged logger (`[LEVEL] [TAG] message`) were filed under the level instead of the component, and printf format strings were never applied, so `%s`/`%d` stayed literal with the values trailing behind them unlabelled — including every LiveWS connection line, where the format is deliberate hardening against injected format specifiers ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10770)).
--- a/changelog.d/fixes/10774-claude-code-flat-rate.md
+++ b/changelog.d/fixes/10774-claude-code-flat-rate.md
@@ -0,0 +1 @@
+- **fix(analytics):** Claude Code (`claude`/`cc`) is a flat-rate subscription, so cost analytics reports `$0` for it instead of estimating Anthropic list prices — the metered `anthropic` API keeps its real cost, and budget/quota/routing still estimate as before ([#10774](https://github.com/diegosouzapw/OmniRoute/pull/10774)) — thanks @electrumguy
--- a/changelog.d/fixes/10781-wal-truncate-scheduler.md
+++ b/changelog.d/fixes/10781-wal-truncate-scheduler.md
@@ -0,0 +1 @@
+- fix(db): periodically run `wal_checkpoint(TRUNCATE)` so the SQLite WAL file shrinks on long-running servers (default 6h, override with `OMNIROUTE_WAL_TRUNCATE_INTERVAL_MS`, `0` disables) (#10781)
--- a/changelog.d/fixes/10782-ws-heartbeat-ping-pong.md
+++ b/changelog.d/fixes/10782-ws-heartbeat-ping-pong.md
@@ -0,0 +1 @@
+- fix(sse): replace LiveWS's application-only liveness check with a protocol-level `ws.ping()`/`pong` heartbeat (RFC 6455 §5.5.2) alongside the existing one, so a read-only dashboard subscriber that never sends anything survives the connection timeout — a socket that stops reading frames entirely is still reaped exactly as before (#10782)
--- a/changelog.d/fixes/10792-double-transport-retry-scope.md
+++ b/changelog.d/fixes/10792-double-transport-retry-scope.md
@@ -0,0 +1 @@
+- **fix(resilience):** scope the same-account transport retry (#9708) out of emergency-fallback and combo hops — it was retrying the free fallback model and combo targets too, doubling upstream calls and corrupting the terminal error status on those paths.
--- a/changelog.d/fixes/9692-openai-to-claude-tool-images.md
+++ b/changelog.d/fixes/9692-openai-to-claude-tool-images.md
@@ -0,0 +1 @@
+- **fix(translator):** convert OpenAI `image_url` blocks nested in `role: "tool"` / `tool_result` content to Claude `image` source blocks so OpenAI-compatible clients (Kimi Code CLI `ReadMediaFile`, and any other tool that returns media) no longer 400 the next Claude-format upstream turn ([#9692](https://github.com/diegosouzapw/OmniRoute/issues/9692))
--- a/changelog.d/fixes/9708-codex-same-account-retry.md
+++ b/changelog.d/fixes/9708-codex-same-account-retry.md
@@ -0,0 +1 @@
+- **fix(resilience):** retry a retryable Codex pre-output 502/503/504/507 once on the same account (2–3s jitter) before cooling the connection, and stop translating that mixed pool into an all-accounts quota `429` ([#9708](https://github.com/diegosouzapw/OmniRoute/issues/9708))
--- a/changelog.d/fixes/PENDING-electron-window-hidden-hostname-bind.md
+++ b/changelog.d/fixes/PENDING-electron-window-hidden-hostname-bind.md
@@ -0,0 +1 @@
+- **fix(electron):** desktop window stays hidden on Windows because the embedded Next.js server binds to the machine hostname instead of loopback ([#PENDING](https://github.com/diegosouzapw/OmniRoute/pull/PENDING))
--- a/changelog.d/fixes/assemble-standalone-cpsync-race.md
+++ b/changelog.d/fixes/assemble-standalone-cpsync-race.md
@@ -0,0 +1 @@
+- fix(build): tolerate a same-realpath symlink or stale-typed dest in the standalone bundle assembler, fixing non-deterministic `ERR_FS_CP_EINVAL`/`ERR_FS_CP_DIR_TO_NON_DIR` crashes under heavy concurrent build I/O
--- a/changelog.d/maintenance/10775-remove-dead-enforce-secrets.md
+++ b/changelog.d/maintenance/10775-remove-dead-enforce-secrets.md
@@ -0,0 +1 @@
+- chore(security): remove the unused `enforceSecrets()` duplicate of the boot secret check and pin the live `enforceWebRuntimeEnv()` wiring with a regression test (#10775)
--- a/changelog.d/maintenance/10779-combo-invocation-docs.md
+++ b/changelog.d/maintenance/10779-combo-invocation-docs.md
@@ -0,0 +1 @@
+- **docs:** Custom combos are only invoked by their exact name in the `model` field — `auto` remains a separate zero-config router, and `openrouter/auto` is a paid OpenRouter product, not an alias ([#10779](https://github.com/diegosouzapw/OmniRoute/pull/10779)) — thanks @maxmad64bis
--- a/changelog.d/maintenance/release-v3850-basereds-stream-utils-20260820.md
+++ b/changelog.d/maintenance/release-v3850-basereds-stream-utils-20260820.md
@@ -0,0 +1 @@
+- **fix(tests):** realign the two `stream-utils` passthrough cases that still asserted the pre-#10017 SSE framing — the event-boundary case declares the OpenAI Responses client format it actually exercises, and the metadata case now pins that surviving lines stay inside one event instead of expecting the `:`/`id:` control lines that #10473 stopped forwarding to every client format.
--- a/config/quality/file-size-baseline.json
+++ b/config/quality/file-size-baseline.json
@@ -442,12 +442,14 @@
    "src/app/(dashboard)/dashboard/providers/[id]/components/modals/AddApiKeyModal.tsx": 1062,
    "src/app/(dashboard)/dashboard/providers/[id]/hooks/useProviderConnections.ts": 1051,
    "src/shared/components/ModelSelectModal.tsx": 1138,
-    "src/shared/constants/providers/apikey/gateways.ts": 1255,
+    "src/shared/constants/providers/apikey/gateways.ts": 1268,
    "open-sse/vendor/codex-chatgpt-web/bridge.ts": 1387,
    "_rebaseline_2026_08_11_v3850_merge_storm_provider_registry": "DRIFT do merge-storm 2026-08-11 (99 PRs mergeados no release/v3.8.50). AddApiKeyModal.tsx (PR #8949 ChatGPT Web provider) e useProviderConnections.ts/ModelSelectModal.tsx (PRs #9011 combo test-all, #9499 image combos) = UI nova legitima acima do cap; gateways.ts = god-file de catalogo de providers que cresceu com PRs #9009/#9421/#9468/#9594 (qualquer split arriscaria corromper o merge de novo — o proprio PR #9421 quebrou o arquivo); bridge.ts (PR #8949) = ponte Chromium vendored; proxyFetch.ts 1207->1220 = drift herdado de merges. Owner autorizou rebaseline com anotacao (2026-08-11).",
    "src/lib/modelCapabilities.ts": 1006,
    "src/app/(dashboard)/dashboard/providers/[id]/providerPageHelpers.ts": 1014,
-    "open-sse/config/imageRegistry.ts": 1019
+    "open-sse/config/imageRegistry.ts": 1019,
+    "src/sse/handlers/chatHelpers.ts": 1017,
+    "src/shared/middleware/chatBodyAdmission.ts": 1005
  },
  "_rebaseline_base_2026_08_10_proxyfetch": "Base-red fix (green-prs sweep, issue #9985): open-sse/utils/proxyFetch.ts 1207 > cap 1000 — new proxied-TLS fetch helper introduced by the Fal reference-image work. Owner-authorized quick rebaseline to green; structural slim tracked for v3.9.0.",
  "_rebaseline_2026_07_27_v3849_train2": "Merge-train 2 (7 PRs) — owner-approved 2026-07-27. Single entry: chatCore.ts 4955->5006 (#8595, Responses multi-turn image compaction before the context hard-reject). Genuine irreducible growth at the existing compaction chokepoint in handleChatCore — the PR adds a last-resort retry against the concrete budget plus the estimateFinalInputTokens helper, both wired at the pre-existing call site rather than a new branch. Covered by tests/unit/8560-responses-image-compaction.test.ts (4 tests).",
@@ -612,5 +614,6 @@
  "_rebaseline_2026_08_12_proxyfetch_redaction": "Base-reds round 3 (#9985): proxyFetch.ts 1220->1239 (+19) = redactProxyDetailsInMessage() helper closing the credential leak #10032 reintroduced (raw proxy URL with user:password appended to the propagated error, Hard Rule #12); irreducible security fix at the existing error-surface chokepoint. Covered by tests/unit/tls-proxy-context.test.ts (strengthened leak guards).",
  "_rebaseline_2026_08_12_modelcapabilities_snapshot_routing": "Base-reds round 3 (#9985): modelCapabilities.ts crossed the new-file cap at 1006 (+~10) when the context/max-input-token override lookups were routed through the #9199 bulk snapshot (fixing 323 per-model SQLite reads per catalog prepare — auto-combo-context-advertising guard); cohesive change at the existing resolution chokepoints, not extractable. Covered by tests/unit/auto-combo-context-advertising.test.ts + model-capability-resolution-snapshot-9199.test.ts.",
  "_rebaseline_2026_08_14_imagetotext_servicekinds": "Image-to-Text category (#10275/#10291): gateways.ts grew 1250→1255 by data lines only — the serviceKinds: [\"llm\", \"imageToText\"] declarations on the openrouter and chutes catalog entries, plus the 3-line comment recording why chutes needs no static dots.ocr entry (passthroughModels discovery). No new logic or branching; the file is a provider catalog of declarative metadata. Splitting a catalog for five lines would be worse than the growth (semantic-families rule).",
-  "_rebaseline_2026_08_18_imageregistry_merge_train": "merge-train 2026-08-18 (owner-authorized, /merge-prs batch of 84): open-sse/config/imageRegistry.ts crossed the 1000-line new-file cap for the first time purely from combining three independent, already-legitimate provider registrations boarded in the same local merge-train — #10542 (aihorde optional-key image catalog), #10494 (gemini-web image generation), #10594 (freepik/magnific provider rename + validation). 996 on release tip -> 1019 on the train tip. Each PR individually adds a small, additive IMAGE_PROVIDERS registry entry at the existing chokepoint; none crosses the cap alone. Not modularized as part of this train's gate fix (out of scope for a merge reconciliation, not a feature change). Covered by each PR's own focused tests (aihorde-image-catalog/generation, gemini-web image tests, freepik/magnific provider tests)."
-}
+  "_rebaseline_2026_08_18_imageregistry_merge_train": "merge-train 2026-08-18 (owner-authorized, /merge-prs batch of 84): open-sse/config/imageRegistry.ts crossed the 1000-line new-file cap for the first time purely from combining three independent, already-legitimate provider registrations boarded in the same local merge-train — #10542 (aihorde optional-key image catalog), #10494 (gemini-web image generation), #10594 (freepik/magnific provider rename + validation). 996 on release tip -> 1019 on the train tip. Each PR individually adds a small, additive IMAGE_PROVIDERS registry entry at the existing chokepoint; none crosses the cap alone. Not modularized as part of this train's gate fix (out of scope for a merge reconciliation, not a feature change). Covered by each PR's own focused tests (aihorde-image-catalog/generation, gemini-web image tests, freepik/magnific provider tests).",
+  "_rebaseline_2026_08_20_v3850_merge_train_batch1": "Merge-train batch1 (2026-08-19/20, 30 PRs boarded onto release/v3.8.50): gateways.ts 1255->1268 = PR #10722 (Token Kiosk OpenAI-compatible provider gateway catalog entry, +13 declarative lines, same god-file no-split rationale as prior gateways.ts rebaselines); chatHelpers.ts (uncapped, not previously frozen) new 1017 = PR #10797 (relay/bifrost error normalization, +23/-2, own-PR growth, existing file already near cap from accumulated chokepoint wiring per its own rebaseline history above); chatBodyAdmission.ts (uncapped) new 1005 = pre-existing base-red on the pure release tip (1004>1000 before this train boarded anything, no PR in this batch touches this file) — frozen here at its current size, not authorizing further growth. Owner-authorized rebaseline (2026-08-19 merge-prs session)."
+}
--- a/config/quality/quality-baseline.json
+++ b/config/quality/quality-baseline.json
@@ -166,7 +166,8 @@
      "dedicatedGate": true
    },
    "zizmorFindings": {
-      "value": 190,
+      "value": 192,
+      "_rebaseline_2026_08_20_radar_export_workflow": "190 -> 192 (+2). Workflow novo `.github/workflows/radar-export.yml` (passo 10 do go-live do Radar: publica o export estável do catálogo como asset de release para o servidor privado baixar via RADAR_EXPORT_URL). Os +2 são unpinned-uses @vN: actions/checkout@v7 + actions/setup-node@v7 — a MESMA convenção deliberada de todos os workflows (ver _scanner_harden_workflows_2026_06_16); fixar por SHA só este violaria a convenção. O findings artipacked do checkout foi CORRIGIDO com `persist-credentials: false` (o job publica via GH_TOKEN em `gh release`, não usa a credencial do checkout). Nenhuma classe nova de template-injection / cache-poisoning / dangerous-triggers. Medido local com zizmor 1.25.2 via `node scripts/check/check-workflows.mjs --ratchet` = 191; +1 do delta conhecido do runner (ver _rebaseline_2026_07_28_ci_runner_delta: o runner enxerga 1 unpinned-uses @vN a mais que o devbox no mesmo commit; a baseline segue o runner) => 192.",
      "_rebaseline_2026_07_20_aliasresolver_hook_split_7808": "175 -> 176 (+1). Companion to PR #7808 (CodeQL js/incomplete-url-substring-sanitization fix in bin/aliasResolver.mjs). The +1 is NOT caused by this PR's code changes (bin/* is not a workflow file) — it is a pre-existing drift that surfaced because the ratchet gate runs on this PR's CI: the zizmor scanner version on the GitHub runner gained a new rule (or extended an existing one) since the v3.8.49 baseline was seeded on 2026-07-17. Breakdown: the new finding is an unpinned-uses @vN class item on one of the existing workflows (same deliberate convention as _scanner_harden_workflows_2026_06_16 — @vN is intentional, SHA-pinning only this one would violate the convention). No new template-injection/artipacked/cache-poisoning/dangerous-triggers classes introduced. Measured by the Quality Gates (Extended) job on run 29713001401 = 176, baseline was 175. Note: by the time this landed on release/v3.8.49, the baseline was already at 176 via _rebaseline_2026_07_17_combo_recovery_hints — this entry is kept as historical record; no further bump applied.",
      "_rebaseline_2026_07_17_v3849_release": "169 -> 175 (+6). Cycle workflow drift (v3.8.48/v3.8.49): npm-publish.yml (new, WS1.3 #7092), electron-release.yml, nightly-compat.yml, nightly-release-green.yml, CI restructures (#7501 full-history base fetch, #7355 main-green, #7202 merge-queue gates, Trunk/Codecov). Breakdown vs v3.8.47: +3 unpinned-uses (@vN convention, deliberate per _scanner_harden_workflows_2026_06_16), +2 cache-poisoning (artifact upload/cache in the OWN electron-release/npm-publish RELEASE workflows -- operator-controlled, not fork-PR exploitable), +1 excessive-permissions (nightly-compat.yml permissions:issues). No new template-injection/artipacked/dangerous-triggers. Measured with zizmor 1.25.2 via `node scripts/check/check-workflows.mjs --ratchet` = 175 on da3a0be69.",
      "direction": "down",
--- a/docs/diagrams/cli-terminal.svg
+++ b/docs/diagrams/cli-terminal.svg
@@ -1,4 +1,4 @@
-<svg viewBox="0 0 1200 350" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="Animated terminal demoing the OmniRoute CLI: omniroute providers list (341 providers registered, anthropic, codex, glm, kimi shown active), omniroute combo list (always-on priority, cost-saver, fusion-panel, context-relay) and omniroute health (healthy, 18412 requests in 24h, p95 412ms, circuit breakers 24 closed, 1 half-open, 0 open), cycling over the 80+ command surface: providers, oauth, keys, combo, nodes, models, cache, compression, cost, usage, quota, health, resilience, telemetry, logs, audit, mcp, a2a, cloud, memory, skills, eval, doctor, repl, tunnel, backup, sync, webhooks, policy, pricing, translator, simulate and more.">
+<svg viewBox="0 0 1200 350" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="Animated terminal demoing the OmniRoute CLI: omniroute providers list (342 providers registered, anthropic, codex, glm, kimi shown active), omniroute combo list (always-on priority, cost-saver, fusion-panel, context-relay) and omniroute health (healthy, 18412 requests in 24h, p95 412ms, circuit breakers 24 closed, 1 half-open, 0 open), cycling over the 80+ command surface: providers, oauth, keys, combo, nodes, models, cache, compression, cost, usage, quota, health, resilience, telemetry, logs, audit, mcp, a2a, cloud, memory, skills, eval, doctor, repl, tunnel, backup, sync, webhooks, policy, pricing, translator, simulate and more.">
 <desc>Compact animated terminal cycling three real OmniRoute CLI commands with a typewriter effect and a scrolling subcommand ticker; the first frame shows the completed providers-list screen.</desc>
 <defs><clipPath id="tickerClip"><rect x="12" y="304" width="1176" height="40"/></clipPath><clipPath id="tw0"><rect x="64" y="46" height="26" width="0"><animate attributeName="width" calcMode="discrete" values="0;31;61;92;122;153;184;214;245;245" keyTimes="0;0.012;0.018;0.024;0.030;0.036;0.042;0.048;0.054;1" dur="18s" repeatCount="indefinite"/></rect></clipPath><clipPath id="tw1"><rect x="64" y="46" height="26" width="0"><animate attributeName="width" calcMode="discrete" values="0;26;51;76;102;128;153;178;204;204" keyTimes="0;0.345;0.351;0.357;0.363;0.369;0.375;0.381;0.387;1" dur="18s" repeatCount="indefinite"/></rect></clipPath><clipPath id="tw2"><rect x="64" y="46" height="26" width="0"><animate attributeName="width" calcMode="discrete" values="0;20;41;61;82;102;122;143;163;163" keyTimes="0;0.678;0.684;0.690;0.696;0.702;0.708;0.714;0.720;1" dur="18s" repeatCount="indefinite"/></rect></clipPath></defs>
 <rect width="1200" height="350" fill="#0d1117"/>
--- a/docs/diagrams/comparison-table.svg
+++ b/docs/diagrams/comparison-table.svg
@@ -1,4 +1,4 @@
-<svg viewBox="0 0 1200 780" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="Comparison table: OmniRoute versus 9router, OpenRouter, CLIProxyAPI and LiteLLM across 13 capabilities. OmniRoute is the only one with the full set: 341 providers, 90+ free providers built-in, 19 routing strategies, 12-engine token compression, a built-in MCP server with 109 tools, A2A protocol, persistent memory, guardrails, cloud agents, TLS fingerprint stealth, desktop/Termux/PWA, 43 UI locales and 100% MIT self-hosted. 9router has free providers, RTK compression and translation but no MCP, A2A, memory, guardrails, cloud agents or stealth. OpenRouter is a hosted SaaS with 400+ models, guardrails and a hosted MCP but is not self-hosted and lacks A2A, memory, cloud agents and stealth. CLIProxyAPI is a light OAuth proxy with two routing strategies. LiteLLM has 100+ providers, A2A and extensive guardrails but no memory, compression, free tier, stealth or cloud agents.">
+<svg viewBox="0 0 1200 780" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="Comparison table: OmniRoute versus 9router, OpenRouter, CLIProxyAPI and LiteLLM across 13 capabilities. OmniRoute is the only one with the full set: 342 providers, 90+ free providers built-in, 19 routing strategies, 12-engine token compression, a built-in MCP server with 109 tools, A2A protocol, persistent memory, guardrails, cloud agents, TLS fingerprint stealth, desktop/Termux/PWA, 43 UI locales and 100% MIT self-hosted. 9router has free providers, RTK compression and translation but no MCP, A2A, memory, guardrails, cloud agents or stealth. OpenRouter is a hosted SaaS with 400+ models, guardrails and a hosted MCP but is not self-hosted and lacks A2A, memory, cloud agents and stealth. CLIProxyAPI is a light OAuth proxy with two routing strategies. LiteLLM has 100+ providers, A2A and extensive guardrails but no memory, compression, free tier, stealth or cloud agents.">
  <desc>Static-header comparison table where each capability row fades in top to bottom; the OmniRoute column is highlighted and shows a check or a leading value in every row, while competitors show a mix of checks, partials and crosses.</desc>
  <defs>
    <pattern id="gC" width="32" height="32" patternUnits="userSpaceOnUse"><path d="M 32 0 L 0 0 0 32" fill="none" stroke="#ffffff" stroke-opacity="0.05" stroke-width="1"/></pattern>
--- a/docs/diagrams/promise-pillars.svg
+++ b/docs/diagrams/promise-pillars.svg
@@ -1,4 +1,4 @@
-<svg viewBox="0 0 1200 540" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="The OmniRoute promise: one endpoint, 341 providers — never stop building, OmniRoute picks the cheapest one that works. Six pillars. Never hit limits: auto-fallback across 341 providers in milliseconds, quota out means the next provider takes over with zero downtime. Save up to 95 percent of tokens: RTK plus Caveman stacked compression cuts 15 to 95 percent of eligible tokens, about 89 percent average on tool-heavy sessions. Zero dollars to start: 90+ providers with a free tier, 56 free forever — Qoder, Pollinations, Cloudflare, SiliconFlow — no card needed. Every tool works: 33 coding agents including Claude Code, Codex, Cursor, Cline, Copilot and Antigravity through one config. One endpoint: OpenAI, Claude, Gemini and Responses API translation — point any tool at /v1 and it just works. Production-grade: circuit breakers, TLS stealth, MCP with 109 tools, A2A, memory, guardrails, evals — 25,000+ tests.">
+<svg viewBox="0 0 1200 540" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="The OmniRoute promise: one endpoint, 342 providers — never stop building, OmniRoute picks the cheapest one that works. Six pillars. Never hit limits: auto-fallback across 342 providers in milliseconds, quota out means the next provider takes over with zero downtime. Save up to 95 percent of tokens: RTK plus Caveman stacked compression cuts 15 to 95 percent of eligible tokens, about 89 percent average on tool-heavy sessions. Zero dollars to start: 90+ providers with a free tier, 56 free forever — Qoder, Pollinations, Cloudflare, SiliconFlow — no card needed. Every tool works: 33 coding agents including Claude Code, Codex, Cursor, Cline, Copilot and Antigravity through one config. One endpoint: OpenAI, Claude, Gemini and Responses API translation — point any tool at /v1 and it just works. Production-grade: circuit breakers, TLS stealth, MCP with 109 tools, A2A, memory, guardrails, evals — 25,000+ tests.">
  <desc>Animated promise card: six pillar tiles fade in in reading order, then a soft colored border highlight sweeps from tile to tile in a continuous cycle.</desc>
  <defs>
    <pattern id="gridPaperP" width="32" height="32" patternUnits="userSpaceOnUse">
@@ -21,7 +21,7 @@
    <line x1="150" y1="53" x2="1160" y2="53" stroke="#232b38" stroke-width="1.5"/>
  </g>
  <g>
-    <text x="40" y="100" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="23" font-weight="600" fill="#c9d1d9">One endpoint. <tspan fill="#a78bfa" font-weight="800">341 providers.</tspan> Never stop building — OmniRoute picks <tspan fill="#7ee787" font-weight="700">the cheapest one that works</tspan>.</text>
+    <text x="40" y="100" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="23" font-weight="600" fill="#c9d1d9">One endpoint. <tspan fill="#a78bfa" font-weight="800">342 providers.</tspan> Never stop building — OmniRoute picks <tspan fill="#7ee787" font-weight="700">the cheapest one that works</tspan>.</text>
  </g>

  <g font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif">
@@ -38,7 +38,7 @@
        <line x1="3.9" y1="3.9" x2="18.1" y2="18.1"/>
      </g>
      <text x="102" y="170" font-size="18" font-weight="800" fill="#74b9ff">Never hit limits</text>
-      <text x="66" y="204" font-size="13.5" fill="#a1a1aa">Auto-fallback across 341 providers in</text>
+      <text x="66" y="204" font-size="13.5" fill="#a1a1aa">Auto-fallback across 342 providers in</text>
      <text x="66" y="226" font-size="13.5" fill="#a1a1aa">milliseconds. Quota out? The next provider</text>
      <text x="66" y="248" font-size="13.5" fill="#a1a1aa">takes over — zero downtime.</text>
    </g>
--- a/docs/diagrams/readme-hero.svg
+++ b/docs/diagrams/readme-hero.svg
@@ -1,4 +1,4 @@
-<svg viewBox="0 0 1200 548" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="OmniRoute hero: Never stop coding. Every AI tool to 341 providers — 90+ free — through one endpoint. Claude Code, Codex, Cursor, Cline, Copilot and Antigravity into free Claude, GPT and Gemini with auto-fallback. RTK + Caveman stacked compression saves 15 to 95 percent of tokens — about 89 percent average on tool-heavy sessions — so you never hit limits. Stats: 341 AI providers, 90+ free tiers, about 1.51B free tokens per month, 15 to 95 percent token savings, 19 routing strategies, zero dollars to start.">
+<svg viewBox="0 0 1200 548" xmlns="http://www.w3.org/2000/svg" role="img" aria-label="OmniRoute hero: Never stop coding. Every AI tool to 342 providers — 90+ free — through one endpoint. Claude Code, Codex, Cursor, Cline, Copilot and Antigravity into free Claude, GPT and Gemini with auto-fallback. RTK + Caveman stacked compression saves 15 to 95 percent of tokens — about 89 percent average on tool-heavy sessions — so you never hit limits. Stats: 342 AI providers, 90+ free tiers, about 1.51B free tokens per month, 15 to 95 percent token savings, 19 routing strategies, zero dollars to start.">
  <desc>Animated hero card: a pulse travels the divider line and a compression bar demo repeatedly shrinks a prompt by up to 95 percent; all headline content is static and readable on the first frame.</desc>
  <defs>
    <pattern id="gridPaperH" width="32" height="32" patternUnits="userSpaceOnUse">
@@ -28,7 +28,7 @@
  <text x="48" y="138" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="60" font-weight="800" fill="#e9edf3">Never stop coding<tspan fill="#a855f7">.</tspan></text>

  <!-- subheadline -->
-  <text x="48" y="184" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="25" font-weight="600" fill="#c9d1d9">Every AI tool → <tspan fill="#a78bfa" font-weight="800">341 providers</tspan> — <tspan fill="#7ee787" font-weight="800">90+ free</tspan> — through one endpoint.</text>
+  <text x="48" y="184" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="25" font-weight="600" fill="#c9d1d9">Every AI tool → <tspan fill="#a78bfa" font-weight="800">342 providers</tspan> — <tspan fill="#7ee787" font-weight="800">90+ free</tspan> — through one endpoint.</text>

  <!-- plug line -->
  <text x="48" y="222" font-family="Inter, 'Segoe UI', Arial, Helvetica, system-ui, sans-serif" font-size="16.5" fill="#a1a1aa">Claude Code · Codex · Cursor · Cline · Copilot · Antigravity&#160;&#160;→&#160;&#160;<tspan fill="#7ee787" font-weight="700">FREE</tspan> Claude / GPT / Gemini · auto-fallback</text>
--- a/docs/getting-started/AUTO-COMBO-GUIDE.md
+++ b/docs/getting-started/AUTO-COMBO-GUIDE.md
@@ -189,7 +189,7 @@ Use `auto/smart` when you want the best quality and are okay with occasional exp

 ### "Can I force a specific provider?"

-Yes! Use a combo with `priority` strategy instead of `auto`. See the [Technical Reference](../routing/AUTO-COMBO.md) for details.
+Yes! Use a combo with `priority` strategy instead of `auto`, then send the combo's **exact name** as the `model` field (e.g. `model: "my-combo"` — not `auto`). See the [Technical Reference](../routing/AUTO-COMBO.md) for details.

 ### "How is this different from round-robin?"

--- a/docs/i18n/ar/llm.txt
+++ b/docs/i18n/ar/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/az/llm.txt
+++ b/docs/i18n/az/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/bg/llm.txt
+++ b/docs/i18n/bg/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/bn/llm.txt
+++ b/docs/i18n/bn/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/cs/llm.txt
+++ b/docs/i18n/cs/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/da/llm.txt
+++ b/docs/i18n/da/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/de/llm.txt
+++ b/docs/i18n/de/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/es/llm.txt
+++ b/docs/i18n/es/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/fa/README.md
+++ b/docs/i18n/fa/README.md
--- a/docs/i18n/fa/llm.txt
+++ b/docs/i18n/fa/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/fi/llm.txt
+++ b/docs/i18n/fi/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/fr/llm.txt
+++ b/docs/i18n/fr/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/gu/llm.txt
+++ b/docs/i18n/gu/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/he/llm.txt
+++ b/docs/i18n/he/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/hi/llm.txt
+++ b/docs/i18n/hi/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/hu/llm.txt
+++ b/docs/i18n/hu/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/id/llm.txt
+++ b/docs/i18n/id/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/in/llm.txt
+++ b/docs/i18n/in/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/it/llm.txt
+++ b/docs/i18n/it/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ja/llm.txt
+++ b/docs/i18n/ja/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ko/llm.txt
+++ b/docs/i18n/ko/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/mr/llm.txt
+++ b/docs/i18n/mr/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ms/llm.txt
+++ b/docs/i18n/ms/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/nl/llm.txt
+++ b/docs/i18n/nl/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/no/llm.txt
+++ b/docs/i18n/no/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/phi/llm.txt
+++ b/docs/i18n/phi/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/pl/llm.txt
+++ b/docs/i18n/pl/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/pt-BR/llm.txt
+++ b/docs/i18n/pt-BR/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/pt/llm.txt
+++ b/docs/i18n/pt/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ro/llm.txt
+++ b/docs/i18n/ro/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ru/llm.txt
+++ b/docs/i18n/ru/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/sk/llm.txt
+++ b/docs/i18n/sk/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/sv/llm.txt
+++ b/docs/i18n/sv/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/sw/llm.txt
+++ b/docs/i18n/sw/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ta/llm.txt
+++ b/docs/i18n/ta/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/te/llm.txt
+++ b/docs/i18n/te/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/th/llm.txt
+++ b/docs/i18n/th/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/tr/llm.txt
+++ b/docs/i18n/tr/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/uk-UA/llm.txt
+++ b/docs/i18n/uk-UA/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/ur/llm.txt
+++ b/docs/i18n/ur/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/vi/llm.txt
+++ b/docs/i18n/vi/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/zh-CN/llm.txt
+++ b/docs/i18n/zh-CN/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/i18n/zh-TW/llm.txt
+++ b/docs/i18n/zh-TW/llm.txt
@@ -4,7 +4,7 @@

 ---

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -169,7 +169,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -281,7 +281,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -479,7 +479,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/docs/providers/CURSOR-API-KEY-AND-CLI.md
+++ b/docs/providers/CURSOR-API-KEY-AND-CLI.md
@@ -0,0 +1,123 @@
+---
+title: "Cursor API provider and the Cursor CLI passthrough"
+version: 3.8.50
+lastUpdated: 2026-08-19
+---
+
+# Cursor API provider and the Cursor CLI passthrough
+
+Two ways to put Cursor behind OmniRoute without an IDE session:
+
+1. **`cursor-api` provider** (card "Cursor API", alias `cua`): an API-key
+   provider that holds a Cursor user API key (`crsr_…`, generated at
+   `https://cursor.com/dashboard/api`). Any OmniRoute client then reaches
+   Cursor models through `/v1/chat/completions` as `cursor-api/<model>` or
+   `cua/<model>`, with the usual quota, fallback and logging layers. The IDE
+   provider (`cursor`, OAuth/IDE session) is unchanged.
+2. **Cursor CLI passthrough**: point the Cursor CLI (`agent`) at OmniRoute so
+   every RPC the CLI makes is authenticated with an OmniRoute API key, forwarded
+   to Cursor with a `cursor-api` connection's credential, and recorded in the
+   Logs page.
+
+## Why the key is exchanged
+
+`api2.cursor.sh` rejects a raw `crsr_…` key as a Bearer token (401). The Cursor
+CLI first POSTs the key to `/auth/exchange_user_api_key` and receives a session
+JWT that expires after one hour; the returned `refreshToken` carries the same
+`exp`, so refreshing means re-exchanging the key.
+`open-sse/services/cursorApiKeyAuth.ts` does that exchange, caches one session
+token per key, re-exchanges five minutes before expiry and drops the cached
+token when Cursor answers 401. `CursorExecutor` calls it right before opening
+the upstream stream for `cursor-api` connections.
+
+## The `cursor-api` provider
+
+Registry: `open-sse/config/providers/registry/cursor/index.ts`
+(`cursor_apiProvider`, `authType: "apikey"`, same `format`, `baseUrl` and
+`models` as `cursor`). Catalog card:
+`src/shared/constants/providers/apikey/specialty-media.ts`. Executor map:
+`open-sse/executors/index.ts` (`"cursor-api"` / `cua` →
+`new CursorExecutor("cursor-api")`).
+
+Dashboard: Providers → Cursor API → Add API key.
+
+REST:
+
+```bash
+curl -sS -X POST http://localhost:20128/api/providers \
+  -H "Content-Type: application/json" \
+  -d '{"provider":"cursor-api","name":"cursor-api-key","apiKey":"crsr_…","priority":1}'
+```
+
+Then:
+
+```bash
+curl -sS http://localhost:20128/v1/chat/completions \
+  -H "Authorization: Bearer <omniroute-api-key>" \
+  -H "Content-Type: application/json" \
+  -d '{"model":"cursor-api/auto","messages":[{"role":"user","content":"say PONG"}]}'
+```
+
+Notes:
+
+- Model listing for `cursor-api` comes from the static Cursor registry (the
+  same list the IDE provider falls back to); no `cursor-agent` install is
+  needed on the OmniRoute host.
+- `POST /api/providers/{id}/refresh-cursor` is for the `cursor` IDE provider
+  only; `cursor-api` connections have no IDE session to renew.
+
+## Cursor CLI passthrough
+
+Route: `src/app/api/cursor-cli/[...path]/route.ts` →
+`open-sse/handlers/cursorCliProxy.ts`. The prefix `/api/cursor-cli/` is
+registered in `src/shared/constants/publicApiRoutes.ts` because the handler
+enforces its own authentication:
+
+| Path                                                                                                                         | Auth expected from the CLI   | What OmniRoute does                                                                                                                                                |
+| ---------------------------------------------------------------------------------------------------------------------------- | ---------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| `POST /auth/exchange_user_api_key`                                                                                           | `Bearer <OmniRoute API key>` | Validates the key, mints a 1h HS256 JWT (signed with `JWT_SECRET`) and returns it                                                                                  |
+| every other path (`/aiserver.v1.*`, `/agent.v1.AgentService/RunSSE`, `/aiserver.v1.BidiService/BidiAppend`, `/v1/traces`, …) | `Bearer <that JWT>`          | Verifies issuer/audience/expiry, picks an active `cursor-api` connection, swaps the Authorization header for the exchanged Cursor token and streams the reply back |
+
+The CLI decodes `exp` from whatever token it receives, so handing it an opaque
+token makes it re-exchange before almost every request; the minted JWT avoids
+that. A 401 from OmniRoute makes the CLI exchange again.
+
+### Setup
+
+1. Create an OmniRoute API key (Dashboard → API keys) and a `cursor-api`
+   connection.
+2. Tell the CLI to use HTTP/1.1 for the agent stream. In
+   `~/.cursor/cli-config.json`:
+
+   ```json
+   { "network": { "useHttp1ForAgent": true } }
+   ```
+
+   Without this the CLI opens the agent turn over HTTP/2 to a separately
+   configured agent host and only the control-plane RPCs go through the
+   endpoint.
+
+3. Run the CLI against OmniRoute:
+
+   ```bash
+   export CURSOR_API_ENDPOINT=http://localhost:20128/api/cursor-cli
+   export CURSOR_API_KEY=<omniroute-api-key>
+   agent -p --trust "Reply with exactly OK"
+   ```
+
+Every hop lands in Logs as provider `cursor-api`, request type `cursor-cli`,
+path `/api/cursor-cli/<rpc>`, attributed to the OmniRoute API key and the
+connection that served it.
+
+### Failure modes
+
+| Situation                                        | Response to the CLI                           |
+| ------------------------------------------------ | --------------------------------------------- |
+| Unknown OmniRoute key and `REQUIRE_API_KEY=true` | 401 `unauthenticated` on exchange             |
+| `REQUIRE_API_KEY=false`                          | anonymous session (mirrors `/v1/*` behaviour) |
+| Expired / foreign / tampered session JWT         | 401, the CLI re-exchanges                     |
+| OmniRoute API key revoked after exchange         | 401 on the next RPC                           |
+| No active `cursor-api` connection                | 503 `unavailable`                             |
+| Cursor rejects the connection's key              | 401 `unauthenticated`, cached session dropped |
+| Upstream unreachable                             | 502 `unavailable` (sanitized message)         |
+| `JWT_SECRET` unset                               | 503 on exchange                               |
--- a/docs/providers/meta.json
+++ b/docs/providers/meta.json
@@ -7,6 +7,7 @@
    "CHATGPT_WEB",
    "AGENTROUTER",
    "ZED-DOCKER",
-    "CURSOR-DOCKER"
+    "CURSOR-DOCKER",
+    "CURSOR-API-KEY-AND-CLI"
  ]
 }
--- a/docs/reference/ENVIRONMENT.md
+++ b/docs/reference/ENVIRONMENT.md
@@ -95,6 +95,7 @@ OmniRoute uses **SQLite** (via `better-sqlite3`) for all persistence. These vari
 | `OMNIROUTE_CRYPT_KEY`                  | _(unset)_            | `src/lib/db/encryption.ts`                            | **Legacy alias** for `STORAGE_ENCRYPTION_KEY`. Accepted as a fallback when the primary variable is absent.                                                                                                                                      |
 | `OMNIROUTE_API_KEY_BASE64`             | _(unset)_            | `src/lib/db/encryption.ts`                            | **Legacy alias** (Base64-encoded form) accepted as a fallback. Decoded automatically before use.                                                                                                                                                |
 | `OMNIROUTE_DB_HEALTHCHECK_INTERVAL_MS` | _(unset)_            | `src/lib/db/core.ts`                                  | Override the periodic SQLite healthcheck interval (ms). When unset, defaults are derived from `NODE_ENV`.                                                                                                                                       |
+| `OMNIROUTE_WAL_TRUNCATE_INTERVAL_MS`   | `21600000` (6h)      | `src/lib/db/core.ts`                                  | Override the periodic `wal_checkpoint(TRUNCATE)` interval (ms). Auto-checkpoint never shrinks the WAL file itself, and a long-running server never closes its DB. `0` disables.                                                                 |
 | `OMNIROUTE_SKIP_DB_HEALTHCHECK`        | `0`                  | `src/lib/db/core.ts`, `src/lib/db/healthCheck.ts`     | Set to `1` to skip the DB healthcheck entirely on startup. Useful for short-lived tasks and integration tests.                                                                                                                                  |
 | `OMNIROUTE_FORCE_DB_HEALTHCHECK`       | `0`                  | `src/lib/db/core.ts`                                  | Set to `1` to force the DB healthcheck loop on, even when it would normally be skipped (e.g., short-lived tasks).                                                                                                                               |
 | `OMNIROUTE_SKIP_POSTINSTALL`           | `0`                  | `scripts/postinstall.mjs`                             | Set to `1` to skip the native-runtime warm-up during `npm install`. Useful in CI/headless installs where sqlite is already built.                                                                                                               |
--- a/docs/reference/PROVIDER_REFERENCE.md
+++ b/docs/reference/PROVIDER_REFERENCE.md
@@ -1,16 +1,16 @@
 ---
 title: "Provider Reference"
 version: 3.8.50
-lastUpdated: 2026-08-18
+lastUpdated: 2026-08-19
 ---

 # Provider Reference

 > **Auto-generated** from `src/shared/constants/providers.ts` — do not edit by hand.
 > Regenerate with: `npm run gen:provider-reference`
-> **Last generated:** 2026-08-18
+> **Last generated:** 2026-08-19

-Total providers: **341**. See category breakdown below.
+Total providers: **342**. See category breakdown below.

 ## Categories

@@ -91,7 +91,7 @@ Use the dashboard at `/dashboard/providers` to enable, configure, and test each
 | `chatgpt-web-codex` | `cgpt-codex` | ChatGPT Web (Codex) | Web cookie | [link](https://chatgpt.com) | Paste the full ChatGPT Cookie header. OmniRoute verifies it in an isolated headless browser profile. | native |
 | `claude-web` | `cw` | Claude Web | Web cookie | [link](https://claude.ai) | Paste your session cookie from claude.ai | none |
 | `conol-web` | `cnl` | Conol (Unofficial/Experimental) | Web cookie | [link](https://conol.ai) | Use browser sign-in, or paste the full Cookie header from conol.ai. The __Secure-better-auth.session_token cookie is required. | — |
-| `copilot-m365-web` | `m365copilot` | Microsoft 365 Copilot (BizChat) | Web cookie | [link](https://m365.cloud.microsoft/chat) | Sign in at m365.cloud.microsoft/chat, then open DevTools → Network → filter 'WS' → click the Chathub WebSocket connection. Copy both the access_token query parameter AND the account-specific Chathub path segment from its request URL (wss://…/Chathub/<path>?…&access_token=…). It is NOT an Authorization: Bearer header on an XHR/Fetch request. The token is short-lived; this is an unofficial integration. | — |
+| `copilot-m365-web` | `m365copilot` | Microsoft 365 Copilot (BizChat) | Web cookie | [link](https://m365.cloud.microsoft/chat) | Sign in at m365.cloud.microsoft/chat, then open DevTools → Network → filter 'WS' → click the Chathub WebSocket connection. Copy both the access_token query parameter AND the account-specific Chathub path segment from its request URL (wss://…/Chathub/<path>?…&access_token=…). It is NOT an Authorization: Bearer header on an XHR/Fetch request. The token is short-lived; this is an unofficial integration. Optional: store a refresh_token in providerSpecificData.refreshToken (any Microsoft device-code/refresh flow for the substrate.office.com/sydney scopes) and OmniRoute pre-flight-refreshes the access token itself — otherwise re-capture after every ~75 min expiry. | — |
 | `copilot-web` | `copilot` | Microsoft Copilot Web | Web cookie | [link](https://copilot.microsoft.com) | Paste the access_token from an authenticated copilot.microsoft.com request (DevTools → Network → Authorization), or export a HAR while logged in | — |
 | `deepseek-web` | `ds-web` | DeepSeek Web | Web cookie | [link](https://chat.deepseek.com) | Paste your userToken from chat.deepseek.com — DevTools → Application → Local Storage → userToken | emulated |
 | `doubao-web` | `db` | Dola Web (ByteDance) | Web cookie | [link](https://www.dola.com) | Paste the full Cookie header from www.dola.com. It should include sessionid, ttwid, and s_v_web_id. If s_v_web_id is unavailable, fp=verify_... from a chat/completion request URL can be used as a fallback. | — |
@@ -120,7 +120,7 @@ Use the dashboard at `/dashboard/providers` to enable, configure, and test each
 | `zai-web` | `zw` | Z.ai Web | Web cookie | [link](https://chat.z.ai) | Copy the "token" value from chat.z.ai → DevTools → Application → Local Storage. Do not copy cookies; OmniRoute handles the per-request CAPTCHA through its browser transport. | — |
 | `zenmux-free` | `zmf` | ZenMux Free (Web) | Web cookie | [link](https://zenmux.ai) | Login at zenmux.ai, then export all cookies using EditThisCookie or Cookie-Editor and paste the full Cookie header string here. Refresh every ~30 days. | — |

-## API Key Providers (paid / paid-with-free-credits) (228)
+## API Key Providers (paid / paid-with-free-credits) (229)

 | ID | Alias | Name | Tags | Website | Notes |
 |----|-------|------|------|---------|-------|
@@ -169,6 +169,7 @@ Use the dashboard at `/dashboard/providers` to enable, configure, and test each
 | `command-code` | `cmd` | Command Code | API key | [link](https://commandcode.ai/) | Use a Command Code API key. Requests are sent to Command Code's /alpha/generate endpoint. |
 | `coze` | `coze` | Coze | API key | [link](https://coze.com) | Get API key at coze.com/open/api |
 | `crof` | `crof` | CrofAI | API key | [link](https://crof.ai) | — |
+| `cursor-api` | `cua` | Cursor API | API key | [link](https://cursor.com/dashboard/api) | Paste a Cursor user API key (crsr_...) from cursor.com/dashboard/api. OmniRoute exchanges it for a session token on demand; no IDE or cursor-agent install is needed. Usage bills to the Cursor plan that owns the key. |
 | `dahl` | `dahl` | Dahl | API key | [link](https://inference.dahl.global) | Click 'Add Account' to auto-generate a token, or add a manual API key. |
 | `databricks` | `databricks` | Databricks | API key, enterprise | [link](https://www.databricks.com) | — |
 | `datarobot` | `datarobot` | DataRobot | API key, enterprise | [link](https://docs.datarobot.com) | Use your DataRobot API token. Optional Base URL can be the account root (for LLM Gateway) or a deployment URL under /api/v2/deployments/<id>. |
@@ -323,7 +324,8 @@ Use the dashboard at `/dashboard/providers` to enable, configure, and test each
 | `tinyfish` | `tf` | TinyFish Fetch | API key | [link](https://docs.tinyfish.ai/fetch-api) | X-API-Key from agent.tinyfish.ai/api-keys |
 | `together` | `together` | Together AI | API key, video | [link](https://www.together.ai) | — |
 | `tokenreply` | `tokenreply` | TokenReply | API key, aggregator | [link](https://www.tokenreply.com) | Free-tagged models have model- and campaign-specific daily limits; no fixed global free quota is published. |
-| `tokenrouter` | `trk` | TokenRouter | API key | [link](https://tokenrouter.com) | Use your TokenRouter API key in Authorization: Bearer <key>. Fully OpenAI-compatible. API base URL: https://api.tokenrouter.com/v1. |
+| `tokenrouter` | `trk` | TokenRouter | API key | [link](https://tokenrouter.com) | Use your TokenRouter API key in Authorization: Bearer *** Fully OpenAI-compatible. API base URL: https://api.tokenrouter.com/v1. |
+| `token-kiosk` | `tk` | Token Kiosk | API key | [link](https://agent-router.gaib.ai) | Use your Token Kiosk API key in Authorization: Bearer *** Fully OpenAI-compatible. API base URL: https://agent-router.gaib.ai/v1. |
 | `topaz` | `topaz` | Topaz | API key, image | [link](https://topazlabs.com) | — |
 | `typhoon` | `typhoon` | Typhoon | API key | [link](https://docs.opentyphoon.ai) | Free API key with a 5 req/s and 200 req/m rate limit. |
 | `udio` | `udio` | Udio | API key | [link](https://udio.com) | Paste session cookie from udio.com (Supabase auth) |
--- a/docs/routing/AUTO-COMBO.md
+++ b/docs/routing/AUTO-COMBO.md
@@ -159,6 +159,28 @@ enumerating every existing combo that shadows a model id, so operators who
 hit this by accident (rather than intentionally, per #6940) have a signal.
 The detection helper lives in `src/lib/combos/modelNameCollision.ts`.

+## Calling a Custom Combo From a Client
+
+Persisted combos (Settings → Combos) are only used when the client sends the combo's **exact name** in the `model` field — there is no fuzzy or partial matching of the combo name, and no `auto/` prefix involved. Resolution order (`getComboForModel()` in `src/sse/services/model.ts`):
+
+1. exact combo-name match (`model: "my-combo"`),
+2. `combo/<name>` prefix (`model: "combo/my-combo"`),
+3. model→combo glob mappings (`/api/model-combo-mappings`).
+
+```bash
+curl -X POST http://localhost:20128/v1/chat/completions \
+  -H "Authorization: Bearer <key>" \
+  -H "Content-Type: application/json" \
+  -d '{"model":"my-combo","messages":[{"role":"user","content":"Hello"}]}'
+```
+
+Two common pitfalls:
+
+- **`auto` does not use your combos.** `auto`/`auto/*` builds its own zero-config candidate pool and only consults persisted combos if a combo is literally named `auto` (not recommended). To route through a combo, send its exact name — not `auto`.
+- **`openrouter/auto` is a real paid OpenRouter product** ("Auto Best Available"), not an OmniRoute alias. It is the single static model entry of the OpenRouter registry (`open-sse/config/providers/registry/openrouter/index.ts`) and is billed separately. Use Settings → Routing → Hide paid models to exclude it from `auto` pools.
+
+See [#7992](https://github.com/diegosouzapw/OmniRoute/issues/7992) and [#7111](https://github.com/diegosouzapw/OmniRoute/issues/7111) for the original confusion this documents.
+
 ## How It Works (Persisted Auto-Combos)

 The Auto-Combo Engine dynamically selects the best provider/model for each request using a **14-factor scoring function** (defined in `open-sse/services/autoCombo/scoring.ts` → `DEFAULT_WEIGHTS`). Weights form a normalized distribution (custom weights are renormalized by `normalizeScoringWeights()`).
--- a/electron/main.js
+++ b/electron/main.js
@@ -781,6 +781,14 @@ function startNextServer() {
      ...serverEnv,
      DATA_DIR: dataDir,
      PORT: String(serverPort),
+      // Pin the embedded server to loopback. Next.js standalone binds to
+      // `process.env.HOSTNAME || '0.0.0.0'`, and Windows always exports
+      // HOSTNAME as the machine name — which resolves to the LAN address, so
+      // the server listens only there and 127.0.0.1 stays closed. The renderer
+      // then fails to load `http://localhost:<port>`, "ready-to-show" never
+      // fires, and the window (created with `show: false`) is never shown.
+      // Mirrors scripts/dev/run-next-playwright.mjs, which already pins this.
+      HOSTNAME: "127.0.0.1",
      NODE_ENV: "production",
      ELECTRON_RUN_AS_NODE: "1",
      NODE_PATH: resolveServerNodePath(serverEnv, resolvePackNodePaths(dataDir)),
--- a/examples/quickstart/python_requests.py
+++ b/examples/quickstart/python_requests.py
@@ -26,3 +26,8 @@ data = {
 response = requests.post(API_URL, headers=headers, json=data)
 response.raise_for_status()
 print(response.json()["choices"][0]["message"]["content"])
+
+# Fresh install, zero credentials — `auto` already works:
+# curl http://localhost:20128/v1/chat/completions \
+#   -H "Content-Type: application/json" \
+#   -d '{"model":"auto","messages":[{"role":"user","content":"Hello!"}]}'
--- a/llm.txt
+++ b/llm.txt
@@ -1,6 +1,6 @@
 # OmniRoute

-> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 341 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.
+> OmniRoute is a free, open-source AI Gateway that acts as a universal API proxy for multi-provider LLMs. It provides smart routing, automatic fallback, load balancing, and format translation across 342 AI providers — all through a single OpenAI-compatible endpoint. Includes a built-in MCP Server (109 tools), A2A v0.3 protocol, Memory/Skills systems, Cloud Agents (codex, cursor, devin, jules), Guardrails framework, and an Electron desktop app.

 ## Overview

@@ -165,7 +165,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 │   │   └── manager.ts            # MITM proxy manager
 │   ├── shared/                   # Shared utilities, components, and constants
 │   │   ├── components/           # Reusable UI components (Card, Badge, Button, Modal, Sidebar, ProviderIcon, etc.)
-│   │   ├── constants/            # Provider definitions (341), model lists, pricing, routing strategies, MCP scopes
+│   │   ├── constants/            # Provider definitions (342), model lists, pricing, routing strategies, MCP scopes
 │   │   ├── contracts/            # Shared API contracts
 │   │   ├── hooks/                # React hooks
 │   │   ├── middleware/           # Shared middleware utilities
@@ -277,7 +277,7 @@ OmniRoute solves the problem of managing multiple AI provider subscriptions, quo
 ## Key Features (v3.8.50)

 ### Core Proxy
- **341 AI providers** with automatic format translation
+- **342 AI providers** with automatic format translation
 - **Provider categories**: Free (90+ free tiers), OAuth, API Key, Self-Hosted, Custom (OpenAI/Anthropic-compatible)
 - **18 routing strategies**: priority, weighted, round-robin, fill-first, p2c, random, least-used, cost-optimized, reset-aware, reset-window, headroom, strict-random, auto, lkgp, context-optimized, context-relay, fusion, pipeline
 - **4-tier fallback**: Subscription → API Key → Cheap → Free
@@ -475,7 +475,7 @@ diagnostics) plus **memory**, **skill**, **agentSkill**, **githubSkill**, **pool

 ## v3.8.x Highlights

- **341-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
+- **342-provider catalog** with 90+ free tiers, one-click account imports, and bulk key add
 - **19 routing strategies** — including `fusion` (parallel panel + judge synthesis), `pipeline`, `reset-aware`, `reset-window`, `headroom`, and `context-relay`
 - **14-factor Auto-Combo scoring** with bandit exploration and progressive cooldown
 - **MCP server expanded to 109 tools / 33 scopes** (canonical + memory/skill/agentSkill/githubSkill/pool/notion/obsidian/localCorpus/gamification/plugin modules)
--- a/open-sse/config/codexClient.ts
+++ b/open-sse/config/codexClient.ts
@@ -1,9 +1,13 @@
 import {
+  CODEX_CLI_RS_ORIGINATOR,
  DEFAULT_CODEX_CLIENT_VERSION,
  getCodexCliRsHeaders as buildCodexCliRsHeaders,
 } from "@/shared/constants/codexClient";

-export { DEFAULT_CODEX_CLIENT_VERSION } from "@/shared/constants/codexClient";
+export {
+  DEFAULT_CODEX_CLIENT_VERSION,
+  CODEX_CLI_RS_ORIGINATOR,
+} from "@/shared/constants/codexClient";
 const DEFAULT_CODEX_USER_AGENT_PLATFORM = "Windows 10.0.26200";
 const DEFAULT_CODEX_USER_AGENT_ARCH = "x64";
 const CODEX_VERSION_OVERRIDE_ENV = "CODEX_CLIENT_VERSION";
@@ -51,6 +55,35 @@ export function getCodexCliRsHeaders(): Record<string, string> {
  return buildCodexCliRsHeaders(getCodexClientVersion());
 }

+/**
+ * Identity for the credential face (auth.openai.com: token exchange / refresh).
+ * The real Codex client sends only `originator` + `User-Agent` on that face
+ * (codex-rs login/default_client.rs default_headers()); the `Version` header
+ * gate exists only on the chatgpt.com/backend-api inference face, so it is
+ * deliberately omitted here. Mirrors sub2api v0.1.178
+ * ApplyCodexCanonicalAuthIdentity.
+ */
+export function getCodexAuthIdentityHeaders(): Record<string, string> {
+  return {
+    "User-Agent": getCodexUserAgent(),
+    originator: CODEX_CLI_RS_ORIGINATOR,
+  };
+}
+
+/**
+ * Canonical Codex CLI identity for server-initiated calls against the
+ * chatgpt.com/backend-api face that are not tied to one end-client request
+ * (usage / quota / models manifest / reset-credits). Same UA/version chain as
+ * inference so these calls do not show up upstream as anonymous half-identities.
+ */
+export function getCodexBackendIdentityHeaders(): Record<string, string> {
+  return {
+    "User-Agent": getCodexUserAgent(),
+    originator: CODEX_CLI_RS_ORIGINATOR,
+    Version: getCodexClientVersion(),
+  };
+}
+
 export function normalizeCodexSessionId(value: unknown): string | null {
  if (typeof value !== "string") return null;
  const normalized = value.trim();
--- a/open-sse/config/codexIdentity.ts
+++ b/open-sse/config/codexIdentity.ts
@@ -1,15 +1,30 @@
 import { createHash, randomUUID } from "node:crypto";

 import { normalizeCodexSessionId } from "./codexClient.ts";
+import { isCrossAccountCodexTurnState, readCodexTurnStateHeader } from "./codexTurnState.ts";

 const CODEX_INSTALLATION_SALT = "omniroute-codex-installation";
 const CODEX_SESSION_SEED_PREFIX = "omniroute:codex-session-id:v1:";
 const CODEX_THREAD_SEED_PREFIX = "omniroute:codex-thread-id:v1:";
+// v2 derivations are keyed by the persisted per-connection random seed
+// (codexFingerprintSeed) instead of the connection-id chain, mirroring
+// sub2api v0.1.178 (#5696): deterministic derivation stays stable, but the
+// seed is generated per connection so identities never collide across
+// deployments and survive connection export/import.
+const CODEX_INSTALLATION_SEED_PREFIX_V2 = "omniroute:codex-installation:v2:";
+const CODEX_SESSION_SEED_PREFIX_V2 = "omniroute:codex-session-id:v2:";
+const CODEX_THREAD_SEED_PREFIX_V2 = "omniroute:codex-thread-id:v2:";
 const UUID_PATTERN = /^[0-9a-f]{8}-[0-9a-f]{4}-[1-5][0-9a-f]{3}-[89ab][0-9a-f]{3}-[0-9a-f]{12}$/i;

 export const CODEX_FINGERPRINT_MODES = ["off", "device", "session", "full"] as const;
 export type CodexFingerprintMode = (typeof CODEX_FINGERPRINT_MODES)[number];
 export const CODEX_FINGERPRINT_MODE_KEY = "codexFingerprintMode";
+/**
+ * System-managed per-connection random seed used as the fingerprint
+ * derivation source. Never sent upstream, stripped from API responses, and
+ * preserved across connection updates (sub2api `codex_fingerprint_seed`).
+ */
+export const CODEX_FINGERPRINT_SEED_KEY = "codexFingerprintSeed";

 export type CodexClientIdentity = {
  mode: CodexFingerprintMode;
@@ -72,6 +87,61 @@ function accountSeed(
  );
 }

+/** The persisted system-managed random seed, when present and a valid UUID. */
+export function getCodexFingerprintSeed(
+  providerSpecificData?: Record<string, unknown> | null
+): string | null {
+  return normalizeUuid(providerSpecificData?.[CODEX_FINGERPRINT_SEED_KEY]);
+}
+
+/** Modes that rewrite account-scoped identifiers and therefore need a stable seed. */
+export function codexFingerprintModeRequiresSeed(mode: CodexFingerprintMode): boolean {
+  return mode === "device" || mode === "session" || mode === "full";
+}
+
+/**
+ * Ensure a Codex OAuth connection carries a persisted fingerprint seed when its
+ * convergence mode derives account-scoped identifiers. Called at connection
+ * create/update time (the persistence layer owns the write); the request path
+ * only ever READS the seed, so an identity never rotates mid-flight.
+ *
+ * Semantics mirror sub2api v0.1.178 `prepareCodexFingerprintExtraFor{Create,Update}`:
+ * - the key is system-managed: any client-supplied value is stripped first;
+ * - an existing valid seed is ALWAYS carried forward (even when the new mode
+ *   is `off` — it stays dormant, ready if convergence is re-enabled later);
+ * - otherwise a fresh seed is created only when the mode requires one
+ *   (device/session/full; the OmniRoute default is session).
+ *
+ * Returns the (possibly new) providerSpecificData, or undefined when there is
+ * nothing to store. Pre-seed connections keep their legacy connection-id
+ * derived identity until the next save — one deliberate rotation, same as
+ * sub2api's migration-225 backfill.
+ */
+export function ensureCodexFingerprintSeed(
+  providerSpecificData?: Record<string, unknown> | null,
+  credentials?: { accessToken?: unknown; refreshToken?: unknown } | null,
+  existingProviderSpecificData?: Record<string, unknown> | null
+): Record<string, unknown> | undefined {
+  const psd: Record<string, unknown> = { ...(providerSpecificData || {}) };
+  // System-managed key: never trust an inbound value, regardless of auth type.
+  delete psd[CODEX_FINGERPRINT_SEED_KEY];
+  if (!isCodexOAuthCredentials(credentials)) {
+    return Object.keys(psd).length > 0 ? psd : undefined;
+  }
+
+  const existingSeed = getCodexFingerprintSeed(existingProviderSpecificData);
+  if (existingSeed) {
+    psd[CODEX_FINGERPRINT_SEED_KEY] = existingSeed;
+    return psd;
+  }
+  const mode = getCodexFingerprintMode(psd, true);
+  if (codexFingerprintModeRequiresSeed(mode)) {
+    psd[CODEX_FINGERPRINT_SEED_KEY] = randomUUID();
+    return psd;
+  }
+  return Object.keys(psd).length > 0 ? psd : undefined;
+}
+
 function readNamedHeader(
  headers: Headers | Record<string, unknown> | null | undefined,
  name: string
@@ -120,6 +190,11 @@ export function getCodexInstallationId(
  const explicit = normalizeUuid(providerSpecificData?.codexInstallationId);
  if (explicit) return explicit;

+  const persistedSeed = getCodexFingerprintSeed(providerSpecificData);
+  if (persistedSeed) {
+    return deriveStableUUIDv4(`${CODEX_INSTALLATION_SEED_PREFIX_V2}${persistedSeed}`);
+  }
+
  const legacyStableSource =
    nonEmptyString(providerSpecificData?.workspaceId) ||
    nonEmptyString(providerSpecificData?.accountId) ||
@@ -137,6 +212,10 @@ export function getCodexConvergedSessionId(
  providerSpecificData?: Record<string, unknown> | null,
  accountKey?: string | null
 ): string {
+  const persistedSeed = getCodexFingerprintSeed(providerSpecificData);
+  if (persistedSeed) {
+    return deriveStableUUIDv4(`${CODEX_SESSION_SEED_PREFIX_V2}${persistedSeed}`);
+  }
  return deriveStableUUIDv4(
    `${CODEX_SESSION_SEED_PREFIX}${accountSeed(providerSpecificData, accountKey)}`
  );
@@ -148,6 +227,10 @@ export function getCodexConvergedThreadId(
  accountKey?: string | null
 ): string {
  if (!nonEmptyString(clientSessionId)) return "";
+  const persistedSeed = getCodexFingerprintSeed(providerSpecificData);
+  if (persistedSeed) {
+    return deriveStableUUIDv4(`${CODEX_THREAD_SEED_PREFIX_V2}${persistedSeed}:${clientSessionId}`);
+  }
  return deriveStableUUIDv4(
    `${CODEX_THREAD_SEED_PREFIX}${accountSeed(providerSpecificData, accountKey)}:${clientSessionId}`
  );
@@ -163,6 +246,26 @@ export function getCodexClientSessionId(
  );
 }

+/**
+ * Decide what to do with the client's `x-codex-turn-state` echo for the
+ * account about to serve this request. The blob is minted per account by the
+ * upstream; replaying another account's blob after failover is a proxy-only
+ * contradiction, so a known cross-account echo is stripped. Same-account or
+ * unknown provenance passes through unchanged (strip only, never inject).
+ * Independent of the fingerprint-convergence mode — account consistency also
+ * applies to explicit `off` / passthrough.
+ */
+export function resolveCodexTurnStateEcho(
+  clientHeaders?: Headers | Record<string, unknown> | null,
+  accountKey?: string | null
+): string | null {
+  const value = readCodexTurnStateHeader(clientHeaders);
+  if (!value) return null;
+  const sessionId = getCodexClientSessionId(clientHeaders);
+  if (sessionId && isCrossAccountCodexTurnState(sessionId, accountKey)) return null;
+  return value;
+}
+
 /**
 * One identity object for every carrier in one upstream turn.
 * accountKey may be the OmniRoute connection id; it is never sent upstream.
@@ -284,13 +387,19 @@ export function withCodexFingerprintCredentials<T extends CodexCredentialIdentit
 ): T {
  const identity = resolveCodexFingerprintIdentity({ credentials, clientHeaders, body });
  const original = resolveCodexOriginalIdentityHeaders({ credentials, clientHeaders });
-  if (!identity && !original) return credentials;
+  // The turn-state echo guard runs for every Codex request (including compact
+  // and explicit-off), unlike the convergence identity above.
+  const turnStateEcho = credentials
+    ? resolveCodexTurnStateEcho(clientHeaders, credentials.connectionId ?? null)
+    : null;
+  if (!identity && !original && !turnStateEcho) return credentials;
  return {
    ...credentials,
    providerSpecificData: {
      ...(credentials.providerSpecificData || {}),
      ...(identity ? { codexClientIdentity: identity } : {}),
      ...(original ? { codexOriginalIdentityHeaders: original } : {}),
+      ...(turnStateEcho ? { codexTurnStateEcho: turnStateEcho } : {}),
    },
  };
 }
--- a/open-sse/config/codexTurnState.ts
+++ b/open-sse/config/codexTurnState.ts
@@ -0,0 +1,144 @@
+/**
+ * codexTurnState.ts — `x-codex-turn-state` relay bookkeeping and the
+ * cross-account echo guard.
+ *
+ * The upstream mints the opaque turn-state blob under the outbound identity
+ * (including the fingerprint-converged installation/session/thread ids), and
+ * the real Codex client echoes it back on later requests of the same turn —
+ * codex-rs captures it from the /responses SSE, the /responses/compact JSON,
+ * and the WS handshake (codex-api/src/sse/responses.rs, endpoint/compact.rs).
+ *
+ * Replaying a blob to the SAME account is self-consistent. Replaying it to a
+ * DIFFERENT account (failover rotated the connection while the client still
+ * echoes the old account's blob) is a contradiction only a proxy chain can
+ * produce — a real Codex client never emits it. The provenance table records
+ * which connection minted the blob a downstream session last received, and
+ * the outbound guard strips echoes known to come from another account.
+ *
+ * Mirrors sub2api v0.1.177 `openai_codex_turn_state.go` (commit 8219dcfc8).
+ * OmniRoute keys the table by the client's original session id only — the
+ * executor pipeline does not carry the API key id, and a real Codex session
+ * id is a random UUID, so accidental cross-key collisions are not a
+ * practical concern.
+ */
+
+const CODEX_TURN_STATE_HEADER = "x-codex-turn-state";
+
+/**
+ * How long a provenance record lives. The blob is echoed within one turn,
+ * but clients may hold it across a whole session; 2h covers the standard
+ * 5-hour quota window's early turns without letting the map grow stale
+ * entries for days.
+ */
+const CODEX_TURN_STATE_TTL_MS = 2 * 60 * 60 * 1000;
+
+/** Opportunistic full sweep every N writes (the read side also lazily expires). */
+const CODEX_TURN_STATE_SWEEP_EVERY_WRITES = 256;
+
+type CodexTurnStateOrigin = {
+  accountKey: string;
+  expiresAt: number;
+};
+
+const turnStateOrigins = new Map<string, CodexTurnStateOrigin>();
+let turnStateWrites = 0;
+
+function normalizeAccountKey(accountKey: unknown): string | null {
+  if (typeof accountKey !== "string") return null;
+  const trimmed = accountKey.trim();
+  return trimmed || null;
+}
+
+/**
+ * Read the turn-state blob from a headers bag (Headers instance or a plain
+ * record with arbitrary casing). Returns null when absent/blank.
+ */
+export function readCodexTurnStateHeader(
+  headers: Headers | Record<string, unknown> | null | undefined
+): string | null {
+  if (!headers) return null;
+  if (headers instanceof Headers) {
+    const value = headers.get(CODEX_TURN_STATE_HEADER);
+    return typeof value === "string" && value.trim() ? value.trim() : null;
+  }
+  if (typeof headers === "object") {
+    for (const [key, value] of Object.entries(headers)) {
+      if (
+        key.toLowerCase() === CODEX_TURN_STATE_HEADER &&
+        typeof value === "string" &&
+        value.trim()
+      ) {
+        return value.trim();
+      }
+    }
+  }
+  return null;
+}
+
+function sweepExpiredTurnStateOrigins(now: number): void {
+  for (const [key, origin] of turnStateOrigins) {
+    if (origin.expiresAt <= now) {
+      turnStateOrigins.delete(key);
+    }
+  }
+}
+
+/**
+ * Record that `accountKey` minted the turn-state blob this downstream session
+ * just received. Must only be called at the response commit point — when the
+ * header is actually written to the client. Recording earlier (e.g. for an
+ * attempt later discarded by failover) would poison the table and make the
+ * guard strip the NEXT account's legitimate echo.
+ */
+export function noteCodexTurnStateProvenance(
+  clientSessionId: string | null | undefined,
+  accountKey: unknown,
+  nowMs?: number
+): void {
+  const sessionId = typeof clientSessionId === "string" ? clientSessionId.trim() : "";
+  const account = normalizeAccountKey(accountKey);
+  if (!sessionId || !account) return;
+
+  const now = typeof nowMs === "number" ? nowMs : Date.now();
+  turnStateOrigins.set(sessionId, {
+    accountKey: account,
+    expiresAt: now + CODEX_TURN_STATE_TTL_MS,
+  });
+
+  turnStateWrites += 1;
+  if (turnStateWrites % CODEX_TURN_STATE_SWEEP_EVERY_WRITES === 0) {
+    sweepExpiredTurnStateOrigins(now);
+  }
+}
+
+/**
+ * Outbound guard: true when the echoed blob is KNOWN to have been minted by a
+ * different account and must be stripped before going upstream. Same-account
+ * or unknown provenance passes through unchanged — stripping only, never
+ * injection (clients that cannot echo are the Claude bridge's concern, not
+ * this module's).
+ */
+export function isCrossAccountCodexTurnState(
+  clientSessionId: string | null | undefined,
+  accountKey: unknown,
+  nowMs?: number
+): boolean {
+  const sessionId = typeof clientSessionId === "string" ? clientSessionId.trim() : "";
+  const account = normalizeAccountKey(accountKey);
+  if (!sessionId || !account) return false;
+
+  const origin = turnStateOrigins.get(sessionId);
+  if (!origin) return false;
+  const now = typeof nowMs === "number" ? nowMs : Date.now();
+  if (origin.expiresAt <= now) {
+    turnStateOrigins.delete(sessionId);
+    return false;
+  }
+  return origin.accountKey !== account;
+}
+
+/** Test hook: forget all provenance records and reset the sweep counter. */
+export function __resetCodexTurnStateOriginsForTesting(): void {
+  turnStateOrigins.clear();
+  turnStateWrites = 0;
+}
--- a/open-sse/config/providers/index.ts
+++ b/open-sse/config/providers/index.ts
@@ -65,7 +65,7 @@ import { api_airforceProvider } from "./registry/api-airforce/index.ts";
 import { mistralProvider } from "./registry/mistral/index.ts";
 import { togetherProvider } from "./registry/together/index.ts";
 import { cohereProvider } from "./registry/cohere/index.ts";
-import { cursorProvider } from "./registry/cursor/index.ts";
+import { cursorProvider, cursor_apiProvider } from "./registry/cursor/index.ts";
 import { volcengineProvider } from "./registry/volcengine/index.ts";
 import { hackclubProvider } from "./registry/hackclub/index.ts";
 import { freetheaiProvider } from "./registry/freetheai/index.ts";
@@ -213,6 +213,7 @@ import { kiroProvider } from "./registry/kiro/index.ts";
 import { openadapterProvider } from "./registry/openadapter/index.ts";
 import { ditProvider } from "./registry/dit/index.ts";
 import { tokenrouterProvider } from "./registry/tokenrouter/index.ts";
+import { token_kioskProvider } from "./registry/token-kiosk/index.ts";
 import { grok_cliProvider } from "./registry/grok-cli/index.ts";
 import { codebuddy_cnProvider } from "./registry/codebuddy-cn/index.ts";
 import { pioneerProvider } from "./registry/pioneer/index.ts";
@@ -326,6 +327,7 @@ export const REGISTRY: Record<string, RegistryEntry> = {
  together: togetherProvider,
  cohere: cohereProvider,
  cursor: cursorProvider,
+  "cursor-api": cursor_apiProvider,
  volcengine: volcengineProvider,
  hackclub: hackclubProvider,
  freetheai: freetheaiProvider,
@@ -475,6 +477,7 @@ export const REGISTRY: Record<string, RegistryEntry> = {
  openadapter: openadapterProvider,
  dit: ditProvider,
  tokenrouter: tokenrouterProvider,
+  "token-kiosk": token_kioskProvider,
  "grok-cli": grok_cliProvider,
  "codebuddy-cn": codebuddy_cnProvider,
  pioneer: pioneerProvider,
--- a/open-sse/config/providers/registry/cursor/index.ts
+++ b/open-sse/config/providers/registry/cursor/index.ts
@@ -156,3 +156,29 @@ export const cursorProvider: RegistryEntry = {
    { id: "kimi-k2.7-code", name: "Kimi K2.7 Code" },
  ],
 };
+
+/**
+ * API-key variant of the Cursor provider.
+ *
+ * Same wire protocol, executor and catalog as `cursor`, but the connection
+ * holds a Cursor user API key (`crsr_…`, cursor.com/dashboard/api) instead of
+ * an IDE/OAuth session. The executor exchanges that key for a session token
+ * on demand (open-sse/services/cursorApiKeyAuth.ts), so no cursor-agent or
+ * IDE install is needed on the OmniRoute host. Kept as a distinct backend ID
+ * so API-key and IDE-session connections never share renewal, quota or
+ * dashboard semantics.
+ */
+export const cursor_apiProvider: RegistryEntry = {
+  id: "cursor-api",
+  alias: "cua",
+  format: cursorProvider.format,
+  executor: "cursor-api",
+  baseUrl: cursorProvider.baseUrl,
+  chatPath: cursorProvider.chatPath,
+  authType: "apikey",
+  authHeader: "bearer",
+  defaultContextLength: cursorProvider.defaultContextLength,
+  headers: getCursorRegistryHeaders(),
+  clientVersion: CURSOR_REGISTRY_VERSION,
+  models: cursorProvider.models,
+};
--- a/open-sse/config/providers/registry/sensenova/index.ts
+++ b/open-sse/config/providers/registry/sensenova/index.ts
@@ -27,6 +27,8 @@ export const sensenovaProvider: RegistryEntry = {
      contextLength: 1048576,
      maxOutputTokens: 65536,
      supportsReasoning: true,
+      supportedThinkingEfforts: ["none", "low", "medium", "high", "xhigh"],
+      supportsXHighEffort: true,
      interleavedField: "reasoning_content",
    },
    {
--- a/open-sse/config/providers/registry/token-kiosk/index.ts
+++ b/open-sse/config/providers/registry/token-kiosk/index.ts
@@ -0,0 +1,20 @@
+import type { RegistryEntry } from "../../shared.ts";
+
+export const token_kioskProvider: RegistryEntry = {
+  id: "token-kiosk",
+  alias: "tk",
+  format: "openai",
+  executor: "default",
+  baseUrl: "https://agent-router.gaib.ai/v1/chat/completions",
+  modelsUrl: "https://agent-router.gaib.ai/v1/models",
+  authType: "apikey",
+  authHeader: "bearer",
+  defaultContextLength: 128000,
+  models: [
+    { id: "claude-3-5-sonnet", name: "Claude 3.5 Sonnet (Token Kiosk)", contextLength: 200000, toolCalling: true, supportsVision: true },
+    { id: "deepseek-v3", name: "DeepSeek V3 (Token Kiosk)", contextLength: 64000, toolCalling: true },
+    { id: "deepseek-r1", name: "DeepSeek R1 (Token Kiosk)", contextLength: 64000, toolCalling: true, supportsReasoning: true },
+    { id: "kimi-k1.5", name: "Kimi K1.5 (Token Kiosk)", contextLength: 128000, toolCalling: true },
+    { id: "minimax-m6", name: "MiniMax M6 (Token Kiosk)", contextLength: 128000, toolCalling: true },
+  ],
+};
--- a/open-sse/executors/base/reasoningEffort.ts
+++ b/open-sse/executors/base/reasoningEffort.ts
@@ -6,6 +6,7 @@ import {
  supportsClaudeMaxEffort,
  supportsXHighEffort,
  getProviderModel,
+  getProviderModels,
 } from "../../config/providerModels.ts";

 /**
@@ -351,6 +352,31 @@ export function sanitizeReasoningEffortForProvider(
  // new models from being unusable for weeks until they're whitelisted (#8057).
  if (effortStr === "max") {
    if (supportsMax) return body; // explicitly known to accept max
+
+    // A model that explicitly advertises its accepted tiers is safe to normalize.
+    // Keep the default pass-through for absent metadata: an unlisted model might
+    // support literal `max`, and #8057 deliberately avoids blocking such models.
+    const providerModelId = modelStr.startsWith(`${provider}/`)
+      ? modelStr.slice(provider.length + 1)
+      : modelStr;
+    // Do not fall back to a globally registered model here. Identical ids can
+    // have different upstream contracts across providers (for example, OpenCode
+    // and SenseNova both expose deepseek-v4-flash with different max support).
+    const explicitEfforts = getProviderModels(provider).find(
+      (entry) => entry.id === providerModelId || entry.aliases?.includes(providerModelId)
+    )?.supportedThinkingEfforts;
+    const maxFallback =
+      Array.isArray(explicitEfforts) && !explicitEfforts.includes("max")
+        ? ["xhigh", "high", "medium", "low"].find((tier) => explicitEfforts.includes(tier))
+        : undefined;
+    if (maxFallback) {
+      log?.info?.(
+        "REASONING_SANITIZE",
+        `${provider}/${modelStr}: downgraded reasoning_effort max → ${maxFallback} (explicit model capability)`
+      );
+      return writeEffortValue(b, maxFallback, c);
+    }
+
    if (!supportsXHigh) {
      // Model is explicitly flagged as rejecting xhigh (and not in supportsMax) —
      // it likely only accepts standard tiers. Degrade to its highest: high.
@@ -360,7 +386,6 @@ export function sanitizeReasoningEffortForProvider(
      );
      return writeEffortValue(b, "high", c);
    }
-    // Default: pass max through unchanged — trust the upstream
    return body;
  }

--- a/open-sse/executors/codex.ts
+++ b/open-sse/executors/codex.ts
@@ -20,6 +20,7 @@ import {
 import { FETCH_BODY_TIMEOUT_MS, HTTP_STATUS, PROVIDERS } from "../config/constants.ts";
 import { readCodexPeekChunk, buildCodexTimeoutSafePassthroughBody } from "./codex/bodyTimeout.ts";
 import {
+  CODEX_CLI_RS_ORIGINATOR,
  getCodexClientVersion,
  getCodexUserAgent,
  normalizeCodexSessionId,
@@ -225,7 +226,6 @@ function convertSystemToDeveloperRole(body: Record<string, unknown>): void {
  }
 }

-
 function stripOrphanedCodexFunctionCallOutputs(body: Record<string, unknown>): void {
  if (!Array.isArray(body.input)) return;
  const input = body.input;
@@ -1045,10 +1045,11 @@ export class CodexExecutor extends BaseExecutor {
      CodexClientIdentity | null | undefined;
    const originalIdentityHeaders = credentials?.providerSpecificData
      ?.codexOriginalIdentityHeaders as Record<string, string> | null | undefined;
+    const turnStateEcho = credentials?.providerSpecificData?.codexTurnStateEcho;

    // Originator header — identifies the client type to the Codex backend.
    // Ref: openai/codex login/src/auth/default_client.rs DEFAULT_ORIGINATOR = "codex_cli_rs"
-    headers["originator"] = "codex_cli_rs";
+    headers["originator"] = CODEX_CLI_RS_ORIGINATOR;

    // session_id header — enables prompt cache affinity on the Codex backend.
    // The official Codex client sets this to conversation_id (a stable UUID per session).
@@ -1060,6 +1061,13 @@ export class CodexExecutor extends BaseExecutor {
    applyCodexOriginalIdentityHeaders(headers, originalIdentityHeaders);
    applyCodexClientIdentityHeaders(headers, clientIdentity);

+    // x-codex-turn-state: forward the client's echo when the provenance guard
+    // (in withCodexFingerprintCredentials) cleared it as same-account. The
+    // blob is account-bound; a stripped (absent) value must stay absent.
+    if (typeof turnStateEcho === "string" && turnStateEcho) {
+      headers["x-codex-turn-state"] = turnStateEcho;
+    }
+
    return headers;
  }

--- a/open-sse/executors/copilot-m365-connection.ts
+++ b/open-sse/executors/copilot-m365-connection.ts
@@ -160,7 +160,16 @@ export function resolveConnectionParams(
  const psd = (credentials?.providerSpecificData ?? {}) as JsonRecord;
  const parsedApiKey =
    typeof credentials?.apiKey === "string" ? parsePastedCredential(credentials.apiKey) : {};
+  // A JWT in credentials.accessToken (3 dot-separated parts — the individual-tier
+  // token is an opaque JWE with 5) is the freshest copy: the executor refreshes it
+  // in place before resolving params, and the framework mutates it after a refresh.
+  const credentialsJwt =
+    typeof credentials?.accessToken === "string" &&
+    credentials.accessToken.split(".").length === 3
+      ? credentials.accessToken
+      : "";
  const accessToken =
+    credentialsJwt ||
    parsedApiKey.accessToken ||
    (typeof credentials?.apiKey === "string" &&
      credentials.apiKey &&
@@ -254,6 +263,135 @@ export function redactWsUrl(wsUrl: string): string {
  return wsUrl.replace(/access_token=[^&]*/i, "access_token=REDACTED");
 }

+// ── OAuth refresh support (#10718 — client ids observed in the browser token
+// and M365-Copilot2API) ────────────────────────────────────────────────────
+//
+// The browser-issued access_token lives ~75 minutes. These helpers redeem a
+// stored refresh_token at the Microsoft identity platform (same public client
+// the m365.cloud.microsoft web app uses) so the connection self-heals instead
+// of requiring a fresh DevTools capture after every expiry.
+
+/** Public client id observed in both the browser token and M365-Copilot2API. */
+export const M365_OAUTH_CLIENT_ID = "c0ab8ce9-e9a0-42e7-b064-33d422df41f1";
+
+export const M365_OAUTH_SCOPE =
+  "openid profile offline_access https://substrate.office.com/sydney/M365Chat.Read " +
+  "https://substrate.office.com/sydney/sydney.readwrite";
+
+/** Refresh lead time — refresh when the current token has less than this left. */
+export const M365_REFRESH_LEAD_MS = 5 * 60 * 1000;
+
+type MinimalLog = {
+  info?: (tag: string, message: string) => void;
+  warn?: (tag: string, message: string) => void;
+};
+
+/** Decode a JWT payload WITHOUT verification — exp/tid are routing hints, never authz. */
+export function decodeJwtClaims(
+  token: string
+): { exp?: number; tid?: string; oid?: string } | null {
+  try {
+    const parts = token.split(".");
+    if (parts.length !== 3) return null;
+    const payload = JSON.parse(Buffer.from(parts[1], "base64url").toString("utf8"));
+    return payload && typeof payload === "object" ? payload : null;
+  } catch {
+    return null;
+  }
+}
+
+/** True when the token is unreadable, already expired, or inside the refresh lead window. */
+export function tokenNeedsRefresh(token: string, leadMs = M365_REFRESH_LEAD_MS): boolean {
+  const claims = decodeJwtClaims(token);
+  if (!claims?.exp) return true;
+  return claims.exp * 1000 <= Date.now() + leadMs;
+}
+
+/** The freshest readable access token for a connection (JWT column → apiKey → psd). */
+export function currentM365AccessToken(
+  credentials: ProviderCredentials | undefined
+): string {
+  if (
+    typeof credentials?.accessToken === "string" &&
+    credentials.accessToken.split(".").length === 3
+  ) {
+    return credentials.accessToken;
+  }
+  if (typeof credentials?.apiKey === "string") {
+    const parsed = parsePastedCredential(credentials.apiKey);
+    if (parsed.accessToken && parsed.accessToken.split(".").length === 3) return parsed.accessToken;
+    // Opaque (JWE) individual-tier token — still a usable credential, just not refreshable.
+    return parsed.accessToken || "";
+  }
+  const psd = (credentials?.providerSpecificData ?? {}) as JsonRecord;
+  if (typeof psd.accessToken === "string") return psd.accessToken;
+  if (typeof psd.access_token === "string") return psd.access_token;
+  return "";
+}
+
+/** The chathub path (`<user-oid>@<tenant-id>`) from wherever it is stored. */
+export function currentM365ChathubPath(credentials: ProviderCredentials | undefined): string {
+  const psd = (credentials?.providerSpecificData ?? {}) as JsonRecord;
+  return (
+    (typeof credentials?.apiKey === "string"
+      ? parsePastedCredential(credentials.apiKey).chathubPath
+      : "") ||
+    (typeof psd.chathubPath === "string" && psd.chathubPath) ||
+    (typeof psd.userTenant === "string" && psd.userTenant) ||
+    ""
+  );
+}
+
+export interface M365RefreshResult {
+  accessToken: string;
+  refreshToken?: string;
+  expiresIn?: number;
+}
+
+/**
+ * Redeem the refresh_token (public client — no secret). MS may rotate the
+ * refresh_token; callers MUST persist the returned one when present or the
+ * token family dies after the first refresh.
+ */
+export async function refreshM365AccessToken(
+  refreshToken: string,
+  tid: string,
+  log?: MinimalLog
+): Promise<M365RefreshResult | { error: string }> {
+  const endpoint = `https://login.microsoftonline.com/${tid || "common"}/oauth2/v2.0/token`;
+  try {
+    const res = await fetch(endpoint, {
+      method: "POST",
+      headers: {
+        "Content-Type": "application/x-www-form-urlencoded",
+        Accept: "application/json",
+      },
+      body: new URLSearchParams({
+        client_id: M365_OAUTH_CLIENT_ID,
+        grant_type: "refresh_token",
+        refresh_token: refreshToken,
+        scope: M365_OAUTH_SCOPE,
+      }),
+    });
+    const data = (await res.json().catch(() => ({}))) as Record<string, unknown>;
+    if (!res.ok || typeof data.access_token !== "string") {
+      const error = typeof data.error === "string" ? data.error : `HTTP ${res.status}`;
+      log?.warn?.("M365_TOKEN", `refresh_token grant failed: ${error}`);
+      return { error };
+    }
+    log?.info?.("M365_TOKEN", "access token refreshed via refresh_token grant");
+    return {
+      accessToken: data.access_token,
+      refreshToken: typeof data.refresh_token === "string" ? data.refresh_token : undefined,
+      expiresIn: typeof data.expires_in === "number" ? data.expires_in : undefined,
+    };
+  } catch (e) {
+    const error = e instanceof Error ? e.message : String(e);
+    log?.warn?.("M365_TOKEN", `refresh request failed: ${error}`);
+    return { error };
+  }
+}
+
 /** Flatten OpenAI messages into a single prompt (system instructions prepended). */
 export function buildPrompt(body: JsonRecord | undefined): string {
  const messages = (body?.messages as Array<JsonRecord>) || [];
--- a/open-sse/executors/copilot-m365-frames.ts
+++ b/open-sse/executors/copilot-m365-frames.ts
@@ -11,7 +11,9 @@
 * Protocol (from @skyzea1's #4042 capture):
 *   - JSON messages terminated with the SignalR record separator `\x1e`.
 *   - Handshake: → {"protocol":"json","version":1}  ← {}  → {"type":6}
- *   - Send: type:4 invocation to target "chat" with arguments[0] = { message, ... }
+ *   - Send: type:4 invocation to target "chat" with arguments[0] = { message, ... },
+ *     immediately followed by a type:1 target:"Metrics" frame in the SAME socket
+ *     write (#10718 — an invocation without its Metrics pair is silently dropped).
 *   - Stream: type:1 target:"update" deltas (bot text at arguments[0].messages[].text,
 *     accumulated — NOT incremental) → isLastUpdate:true → type:2 final → type:3 completion.
 */
@@ -25,19 +27,18 @@ export const HANDSHAKE_REQUEST = { protocol: "json", version: 1 } as const;
 /** SignalR keepalive ping frame. */
 export const KEEPALIVE_PING = { type: 6 } as const;

-/** Allowed message types observed in the individual M365 send frame. */
+/**
+ * Allowed message types observed in the 2026-08 recapture of the working
+ * `m365.cloud.microsoft/chat` client (#10718). The old 11-entry list is no longer
+ * seen on the wire — the stale shape gets closed immediately after the type:4.
+ */
 export const ALLOWED_MESSAGE_TYPES = [
  "Chat",
  "Suggestion",
-  "InternalSearchQuery",
  "Disengaged",
-  "InternalLoaderMessage",
  "Progress",
-  "GeneratedCode",
-  "RenderCardRequest",
-  "AdsQuery",
-  "SemanticSerp",
-  "GenerateContentQuery",
+  "EndOfRequest",
+  "InternalLoaderMessage",
 ] as const;

 /**
@@ -74,22 +75,20 @@ export const M365_ENTERPRISE_EXTRA_MESSAGE_TYPES = [
  "SwitchRespondingEndpoint",
 ] as const;

+/**
+ * Individual / EDU option sets from the 2026-08 recapture (#10718) — 14 entries.
+ * The previous 25-entry consumer/MSA set (enable_msa_user, pdnascan, cwc_code_*,
+ * …) is no longer observed on the wire and belongs to the shape the substrate
+ * now drops silently.
+ */
 export const M365_DEFAULT_OPTION_SETS = [
  "search_result_progress_messages_with_search_queries",
  "update_textdoc_response_after_streaming",
  "deepleo_networking_timeout_10minutes_canmore",
  "cwc_flux_image",
-  "cwc_code_interpreter",
-  "cwc_code_interpreter_amsfix",
-  "enable_msa_user",
-  "cwcgptv",
+  "cwcfluxgptv",
  "flux_v3_gptv_enable_upload_multi_image_in_turn_wo_ch",
  "gptvnorm2048",
-  "pdnascan",
-  "cwc_code_interpreter_citation_fix",
-  "code_interpreter_interactive_charts",
-  "cwc_code_interpreter_interactive_charts_inline_image",
-  "code_interpreter_matplotlib_patching",
  "cwc_fileupload_odb",
  "update_memory_plugin",
  "add_custom_instructions",
@@ -97,9 +96,6 @@ export const M365_DEFAULT_OPTION_SETS = [
  "flux_v3_progress_messages",
  "enable_batch_token_processing",
  "enable_gg_gpt",
-  "flux_v3_image_gen_enable_non_watermarked_storage",
-  "flux_v3_image_gen_enable_story",
-  "rich_responses",
 ] as const;

 /** Append the record separator to a JSON-serializable frame. */
@@ -117,6 +113,32 @@ export function keepaliveFrame(): string {
  return encodeFrame(KEEPALIVE_PING);
 }

+/**
+ * #10718 — the browser follows the type:4 chat invocation with this type:1
+ * target:"Metrics" frame in the SAME socket write. Sending the invocation alone
+ * gets it silently ignored (no update frames at all), so the executor must
+ * concatenate `metricsFrame()` onto the invocation payload.
+ */
+export const CHAT_METRICS_FRAME = {
+  arguments: [
+    {
+      Timestamps: {
+        ConnectionEstablished: "",
+        ConnectionStart: "",
+        UserInputStart: "",
+        UserInputSubmit: "",
+      },
+    },
+  ],
+  target: "Metrics",
+  type: 1,
+} as const;
+
+/** Serialized Metrics follow-up frame (see {@link CHAT_METRICS_FRAME}). */
+export function metricsFrame(): string {
+  return encodeFrame(CHAT_METRICS_FRAME);
+}
+
 /**
 * Split a raw socket buffer into complete `\x1e`-terminated frames, returning any
 * trailing partial frame as `rest` so it can be prepended to the next chunk.
@@ -155,22 +177,37 @@ export function handshakeError(frame: Record<string, unknown> | null): string |

 export interface ChatInvocationOptions {
  text: string;
-  /** Per-connection trace id (hex), reused as clientCorrelationId/traceId. */
+  /** Per-invocation trace id (GUID). */
  traceId: string;
-  /** Per-session id (GUID). */
+  /** Client correlation id; defaults to {@link ChatInvocationOptions.traceId}. */
+  clientCorrelationId?: string;
+  /** Per-session id (GUID, == the WS URL X-SessionId query). */
  sessionId: string;
+  /** Per-request id (== the WS URL chatsessionid/clientrequestid query). */
+  requestId: string;
+  /**
+   * Conversation id — MUST match the ConversationId query of the WS URL the
+   * invocation rides on (#10718: the server cross-checks the two).
+   */
+  conversationId: string;
+  /** BCP-47 locale echoed in message.locale; defaults to "en-us". */
+  locale?: string;
+  /** IANA time zone for message.locationInfo; defaults to "UTC". */
+  timeZone?: string;
+  /** Hour offset for message.locationInfo; defaults to 0. */
+  timeZoneOffset?: number;
  /** Whether this is the first turn of the conversation. */
  isStartOfSession?: boolean;
-  /** Tier-specific option flags; left empty by default (tuned during live validation). */
+  /** Tier-specific option flags; defaults to {@link M365_DEFAULT_OPTION_SETS}. */
  optionsSets?: string[];
  tone?: string;
  /** Tier-specific allowed message types; defaults to {@link ALLOWED_MESSAGE_TYPES}. */
  allowedMessageTypes?: readonly string[];
  /**
-   * Tier-specific disconnect behavior sent in every type:4 chat invocation. The work
-   * Surface rejects any value other than exactly "continue" (#8971). Defaults to ""
-   * for individual/consumer/EDU tiers; {@link resolveChatInvocationOverrides} returns
-   * "continue" for the enterprise tier.
+   * Tier-specific disconnect behavior sent in the type:4 chat invocation. The work
+   * surface rejects any value other than exactly "continue" (#8971), so the
+   * enterprise tier sends it; the 2026-08 recapture shows the individual/EDU
+   * surface omits the key entirely, so it is left out unless set (#10718).
   */
  disconnectBehavior?: string;
 }
@@ -185,7 +222,7 @@ export function resolveChatInvocationOverrides(tier: string | undefined): {
  optionsSets: string[];
  tone: string;
  allowedMessageTypes: readonly string[];
-  disconnectBehavior: string;
+  disconnectBehavior: string | undefined;
 } {
  if (tier === "enterprise") {
    return {
@@ -197,9 +234,12 @@ export function resolveChatInvocationOverrides(tier: string | undefined): {
  }
  return {
    optionsSets: [...M365_DEFAULT_OPTION_SETS],
-    tone: "",
+    // #10718 — the 2026-08 recapture sends tone:"magic" (lowercase) on the
+    // individual/EDU surface; the old "" default is part of the dropped shape.
+    tone: "magic",
    allowedMessageTypes: ALLOWED_MESSAGE_TYPES,
-    disconnectBehavior: "",
+    // Omitted entirely on the individual/EDU wire (see ChatInvocationOptions).
+    disconnectBehavior: undefined,
  };
 }

@@ -207,7 +247,7 @@ export function resolveChatInvocationOverrides(tier: string | undefined): {
 * BizChat exposes several models selected by the `tone` field of the `type:4` chat
 * invocation (#7872, values confirmed against a real enterprise tenant in #7850). Each
 * tone-selected variant is registered as its own model id; the bare `copilot-m365` id is
- * intentionally absent here so it keeps the tier default tone (`Magic` on enterprise, `""`
+ * intentionally absent here so it keeps the tier default tone (`Magic` on enterprise, `magic`
 * otherwise) resolved by {@link resolveChatInvocationOverrides}.
 */
 export const M365_MODEL_TONE_MAP: Readonly<Record<string, string>> = {
@@ -228,7 +268,14 @@ export function resolveToneForModel(model: string | undefined): string | undefin

 /**
 * Build the `type:4` chat invocation frame body (not yet `\x1e`-terminated).
- * Mirrors the argument shape captured on the individual M365 path in #4042.
+ * Mirrors the argument shape recaptured from a working `m365.cloud.microsoft/chat`
+ * client in 2026-08 (#10718). Notable differences from the pre-#10718 shape: a
+ * populated `clientInfo` + `productThreadType:"Office"`, a `conversationId`
+ * matching the WS URL query, a rich `message` object, and no
+ * `spokenTextMode` / `extraExtensionParameters` / `isSbsSupported` /
+ * `renderReferencesBehindEOS` / `disconnectBehavior` — none of those are still
+ * observed on the wire, and the stale shape gets closed immediately after the
+ * invocation.
 */
 export function buildChatInvocation(opts: ChatInvocationOptions): Record<string, unknown> {
  return {
@@ -237,33 +284,48 @@ export function buildChatInvocation(opts: ChatInvocationOptions): Record<string,
    invocationId: "0",
    arguments: [
      {
-        source: "officeweb",
-        clientCorrelationId: opts.traceId,
-        sessionId: opts.sessionId,
-        optionsSets: opts.optionsSets ?? [...M365_DEFAULT_OPTION_SETS],
-        streamingMode: "ConciseWithPadding",
-        spokenTextMode: "None",
-        options: {},
-        extraExtensionParameters: {},
        allowedMessageTypes: opts.allowedMessageTypes
          ? [...opts.allowedMessageTypes]
          : [...ALLOWED_MESSAGE_TYPES],
-        sliceIds: [],
-        threadLevelGptId: {},
-        traceId: opts.traceId,
-        isStartOfSession: opts.isStartOfSession ?? true,
-        clientInfo: {},
-        message: {
-          author: "user",
-          inputMethod: "Keyboard",
-          text: opts.text,
-          messageType: "Chat",
+        clientCorrelationId: opts.clientCorrelationId ?? opts.traceId,
+        clientInfo: {
+          clientAppName: "Office",
+          clientPlatform: "mcmcopilot-web",
        },
+        conversationId: opts.conversationId,
+        isStartOfSession: opts.isStartOfSession ?? true,
+        message: {
+          adaptiveCards: [],
+          attachments: null,
+          author: "user",
+          clientPreferences: {},
+          entityAnnotationTypes: ["People", "File", "Event", "Email", "TeamsMessage"],
+          experienceType: "Default",
+          inputMethod: "Keyboard",
+          locale: opts.locale ?? "en-us",
+          locationInfo: {
+            timeZone: opts.timeZone ?? "UTC",
+            timeZoneOffset: opts.timeZoneOffset ?? 0,
+          },
+          messageType: "Chat",
+          requestId: opts.requestId,
+          text: opts.text,
+        },
+        options: {},
+        optionsSets: opts.optionsSets ?? [...M365_DEFAULT_OPTION_SETS],
        plugins: [],
-        isSbsSupported: false,
-        tone: opts.tone ?? "",
-        renderReferencesBehindEOS: true,
-        disconnectBehavior: opts.disconnectBehavior ?? "",
+        productThreadType: "Office",
+        sessionId: opts.sessionId,
+        sliceIds: [],
+        source: "officeweb",
+        streamingMode: "ConciseWithPadding",
+        threadLevelGptId: {},
+        tone: opts.tone ?? "magic",
+        toolChoice: null,
+        traceId: opts.traceId,
+        // #8971 keeps "continue" for the enterprise tier; the individual/EDU wire
+        // omits the key, so only include it when actually set (#10718).
+        ...(opts.disconnectBehavior ? { disconnectBehavior: opts.disconnectBehavior } : {}),
      },
    ],
  };
--- a/open-sse/executors/copilot-m365-web.ts
+++ b/open-sse/executors/copilot-m365-web.ts
@@ -5,8 +5,13 @@ import { BaseExecutor, type ExecuteInput, type ExecutorLog } from "./base.ts";
 import {
  buildPrompt,
  buildWsUrl,
+  currentM365AccessToken,
+  currentM365ChathubPath,
+  decodeJwtClaims,
  redactWsUrl,
+  refreshM365AccessToken,
  resolveConnectionParams,
+  tokenNeedsRefresh,
 } from "./copilot-m365-connection.ts";
 import {
  accumulateBotContent,
@@ -17,7 +22,7 @@ import {
  handshakeFrame,
  isCompletionFrame,
  isUpdateFrame,
-  keepaliveFrame,
+  metricsFrame,
  parseFrame,
  resolveChatInvocationOverrides,
  resolveToneForModel,
@@ -152,8 +157,17 @@ export class CopilotM365WebExecutor extends BaseExecutor {

          try {
            const wsUrlParts = new URL(input.wsUrl);
-            const traceId = wsUrlParts.searchParams.get("clientrequestid") ?? crypto.randomUUID().replace(/-/g, "");
+            // #10718 — the invocation must echo the ids riding in the WS URL query
+            // (conversationId is cross-checked server-side). traceId is a fresh GUID
+            // per turn, as in the browser capture.
+            const requestId =
+              wsUrlParts.searchParams.get("chatsessionid") ??
+              wsUrlParts.searchParams.get("clientrequestid") ??
+              crypto.randomUUID();
            const sessionId = wsUrlParts.searchParams.get("X-SessionId") ?? crypto.randomUUID();
+            const conversationId =
+              wsUrlParts.searchParams.get("ConversationId") ?? crypto.randomUUID();
+            const traceId = crypto.randomUUID();

            log?.debug?.("M365_WS", `connecting → ${redactWsUrl(input.wsUrl)}`);

@@ -166,23 +180,26 @@ export class CopilotM365WebExecutor extends BaseExecutor {
            });

            const sendChat = () => {
-              ws?.send(keepaliveFrame());
              const overrides = resolveChatInvocationOverrides(input.tier);
              // Model-driven tone (#7872) wins over the tier default; a bare/unknown id
              // keeps the tier tone resolved above.
              const tone = resolveToneForModel(input.model) ?? overrides.tone;
-              ws?.send(
-                encodeFrame(
-                  buildChatInvocation({
-                    text: input.prompt,
-                    traceId,
-                    sessionId,
-                    isStartOfSession: true,
-                    ...overrides,
-                    tone,
-                  })
-                )
+              const invocationFrame = encodeFrame(
+                buildChatInvocation({
+                  text: input.prompt,
+                  traceId,
+                  sessionId,
+                  requestId,
+                  conversationId,
+                  isStartOfSession: true,
+                  ...overrides,
+                  tone,
+                })
              );
+              // #10718 — the invocation and its type:1 Metrics follow-up must land
+              // in ONE socket write, exactly as the browser sends them; a bare
+              // invocation (or one preceded by a type:6 ping) is silently dropped.
+              ws?.send(invocationFrame + metricsFrame());
            };

            ws.on("open", () => {
@@ -273,6 +290,61 @@ export class CopilotM365WebExecutor extends BaseExecutor {
    );
  }

+  /**
+   * #10718 — proactively refresh the M365 access token before opening the WS.
+   * A WS-handshake 401 surfaces as an error event INSIDE the SSE stream (the HTTP
+   * response is already 200 by then), so chatCore's generic 401→refresh→retry
+   * orchestration never triggers — the refresh has to happen here, pre-flight.
+   * No-ops for legacy connections without a stored refresh_token.
+   */
+  private async ensureFreshCredentials(
+    credentials: ExecuteInput["credentials"],
+    onCredentialsRefreshed: ExecuteInput["onCredentialsRefreshed"],
+    log: ExecutorLog | null
+  ): Promise<void> {
+    const psd = (credentials?.providerSpecificData ?? {}) as JsonRecord;
+    const refreshToken =
+      credentials.refreshToken || (typeof psd.refreshToken === "string" ? psd.refreshToken : "");
+    if (!refreshToken) return;
+
+    const current = currentM365AccessToken(credentials);
+    if (current && !tokenNeedsRefresh(current)) return;
+
+    const tid =
+      decodeJwtClaims(current)?.tid || (typeof psd.tid === "string" ? psd.tid : "") || "";
+    const result = await refreshM365AccessToken(refreshToken, tid, log ?? undefined);
+    if ("error" in result) {
+      // Fall through with the existing token — the WS layer will surface the failure.
+      return;
+    }
+
+    const rotated = result.refreshToken || refreshToken;
+    const chathubPath = currentM365ChathubPath(credentials);
+    const next = {
+      ...credentials,
+      accessToken: result.accessToken,
+      refreshToken: rotated,
+      // Keep the pasted-format apiKey self-consistent so every resolution path
+      // (fresh column, stale column, dashboard re-read) sees the same token.
+      ...(chathubPath
+        ? { apiKey: `access_token=${result.accessToken}; chathubPath=${chathubPath}` }
+        : {}),
+      ...(result.expiresIn
+        ? { expiresAt: new Date(Date.now() + result.expiresIn * 1000).toISOString() }
+        : {}),
+    };
+    Object.assign(credentials, next);
+    try {
+      await onCredentialsRefreshed?.(next);
+    } catch (err) {
+      // #7676 pattern: a persistence failure must never fail the user-facing response.
+      log?.warn?.(
+        "M365_TOKEN",
+        `persisting refreshed token failed (${err instanceof Error ? err.message : String(err)}) — will re-refresh next request`
+      );
+    }
+  }
+
  async execute(input: ExecuteInput): Promise<{
    response: Response;
    url: string;
@@ -293,6 +365,12 @@ export class CopilotM365WebExecutor extends BaseExecutor {
      };
    }

+    await this.ensureFreshCredentials(
+      input.credentials,
+      input.onCredentialsRefreshed,
+      input.log ?? null
+    );
+
    const connectionParams = resolveConnectionParams(input.credentials);
    if ("error" in connectionParams) {
      return {
--- a/open-sse/executors/cursor.ts
+++ b/open-sse/executors/cursor.ts
@@ -57,6 +57,13 @@ import {
  type StreamingState as ComposerStreamingState,
 } from "../utils/composerToolCalls.ts";
 import { cursorSessionManager, type CursorSession } from "../services/cursorSessionManager.ts";
+import {
+  CursorApiKeyExchangeError,
+  invalidateCursorSessionToken,
+  isCursorApiKey,
+  resolveCursorBearerToken,
+  stripCursorOAuthTokenPrefix,
+} from "../services/cursorApiKeyAuth.ts";
 import crypto from "crypto";
 import * as fs from "node:fs";
 import * as zlib from "node:zlib";
@@ -706,18 +713,44 @@ export function processFrame(
 }

 export class CursorExecutor extends BaseExecutor {
-  constructor() {
-    super("cursor", PROVIDERS.cursor);
+  constructor(provider: "cursor" | "cursor-api" = "cursor") {
+    super(provider, PROVIDERS[provider]);
  }

  buildUrl() {
    return CURSOR_AGENT_URL;
  }

+  /**
+   * API-key connections carry a `crsr_…` key that api2.cursor.sh does not
+   * accept as a Bearer; swap it for the exchanged session token before the
+   * h2 stream is opened. OAuth/IDE-session connections pass through untouched.
+   */
+  async resolveExecutionCredentials(credentials) {
+    if (!isCursorApiKey(credentials?.apiKey)) return credentials;
+    try {
+      const accessToken = await resolveCursorBearerToken(credentials);
+      return { ...credentials, accessToken };
+    } catch (err) {
+      const status =
+        err instanceof CursorApiKeyExchangeError ? err.status : HTTP_STATUS.SERVER_ERROR;
+      const message = err instanceof Error ? err.message : String(err);
+      return new Response(
+        JSON.stringify({
+          error: {
+            message: sanitizeErrorMessage(message),
+            type: status === HTTP_STATUS.UNAUTHORIZED ? "authentication_error" : "connection_error",
+            code: "",
+          },
+        }),
+        { status, headers: { "Content-Type": "application/json" } }
+      );
+    }
+  }
+
  buildHeaders(credentials) {
-    const accessToken = credentials.accessToken;
    const ghostMode = credentials.providerSpecificData?.ghostMode !== false;
-    const cleanToken = accessToken.includes("::") ? accessToken.split("::")[1] : accessToken;
+    const cleanToken = stripCursorOAuthTokenPrefix(credentials.accessToken ?? "");
    const requestId = crypto.randomUUID();
    const traceParent = `00-${crypto.randomBytes(16).toString("hex")}-${crypto.randomBytes(8).toString("hex")}-01`;

@@ -825,7 +858,7 @@ export class CursorExecutor extends BaseExecutor {
   */
  private async loadLiveCatalogIds(): Promise<ReadonlySet<string> | undefined> {
    try {
-      const catalog = await getActiveSyncedCatalog("cursor");
+      const catalog = await getActiveSyncedCatalog(this.provider);
      if (!catalog.models.length) return undefined;
      return new Set(catalog.models.map((model) => model.id));
    } catch {
@@ -1179,7 +1212,11 @@ export class CursorExecutor extends BaseExecutor {

  async execute({ model, body, stream, credentials, signal, log, upstreamExtraHeaders }) {
    const url = this.buildUrl();
-    const headers = this.buildHeaders(credentials);
+    const executionCredentials = await this.resolveExecutionCredentials(credentials);
+    if (executionCredentials instanceof Response) {
+      return { response: executionCredentials, url, headers: {}, transformedBody: body };
+    }
+    const headers = this.buildHeaders(executionCredentials);
    mergeUpstreamExtraHeaders(headers, upstreamExtraHeaders);

    const messages: ChatMessage[] = body.messages || [];
@@ -1252,8 +1289,10 @@ export class CursorExecutor extends BaseExecutor {
    if (isToolFollowUp) {
      session = cursorSessionManager.acquire(conversationId);
      // #9029: content-based session match when client lacks conversation_id.
-      if (!session && !body.conversation_id) session = cursorSessionManager.findByToolCallIds(
-        messages.filter(m => m.role === "tool" && m.tool_call_id).map(m => m.tool_call_id!));
+      if (!session && !body.conversation_id)
+        session = cursorSessionManager.findByToolCallIds(
+          messages.filter((m) => m.role === "tool" && m.tool_call_id).map((m) => m.tool_call_id!)
+        );
    }

    if (session) {
@@ -1334,6 +1373,9 @@ export class CursorExecutor extends BaseExecutor {
      if (opened.status !== 200) {
        const errBuf = await opened.consumeError();
        const errText = errBuf.toString("utf8") || "Unknown error";
+        if (opened.status === HTTP_STATUS.UNAUTHORIZED && isCursorApiKey(credentials.apiKey)) {
+          invalidateCursorSessionToken(credentials.apiKey);
+        }
        return {
          response: buildErrorResponse(opened.status, `[${opened.status}]: ${errText}`),
          url,
--- a/open-sse/executors/index.ts
+++ b/open-sse/executors/index.ts
@@ -1,9 +1,5 @@
 import { SEARCH_PROVIDERS } from "../config/searchRegistry.ts";
-import {
-  registerExecutor,
-  getRegisteredExecutor,
-  hasRegisteredExecutor,
-} from "./registry.ts";
+import { registerExecutor, getRegisteredExecutor, hasRegisteredExecutor } from "./registry.ts";
 import { AntigravityExecutor } from "./antigravity.ts";
 import { GithubExecutor } from "./github.ts";
 import { GheCopilotExecutor } from "./ghe-copilot.ts";
@@ -107,6 +103,8 @@ const executors = {
  "glm-cn": new GlmExecutor("glm-cn"),
  glmt: new GlmExecutor("glmt"),
  cu: new CursorExecutor(), // Alias for cursor
+  "cursor-api": new CursorExecutor("cursor-api"),
+  cua: new CursorExecutor("cursor-api"),
  "azure-openai": new AzureOpenAIExecutor(),
  "azure-ai": new AzureAiExecutor(),
  "command-code": new CommandCodeExecutor(),
--- a/open-sse/handlers/chatCore.ts
+++ b/open-sse/handlers/chatCore.ts
@@ -41,7 +41,11 @@ import {
  isStripReasoningRequested,
 } from "./chatCore/headers.ts";
 import { markCodexScopeRateLimited } from "./chatCore/codexFailover.ts";
-import { isCodexOriginatedHeaders } from "../config/codexIdentity.ts";
+import { getCodexClientSessionId, isCodexOriginatedHeaders } from "../config/codexIdentity.ts";
+import {
+  noteCodexTurnStateProvenance,
+  readCodexTurnStateHeader,
+} from "../config/codexTurnState.ts";
 import { trackDevice, extractIpFromHeaders } from "../services/deviceTracker.ts";
 import { getCombosCached } from "./chatCore/comboContextCache.ts";
 export { clearCombosCache, clearUpstreamProxyConfigCache } from "./chatCore/comboContextCache.ts";
@@ -722,12 +726,13 @@ export async function handleChatCore({
    copilotCompatibleReasoning,
    clientResponseFormat,
  } = resolveChatCoreRequestFormat({ clientRawRequest, body, provider, userAgent });
-  const nativeOpenAICompatibleResponsesPassthrough = shouldUseNativeOpenAICompatibleResponsesPassthrough({
-    provider,
-    sourceFormat,
-    endpointPath,
-    providerSpecificData: credentials?.providerSpecificData,
-  });
+  const nativeOpenAICompatibleResponsesPassthrough =
+    shouldUseNativeOpenAICompatibleResponsesPassthrough({
+      provider,
+      sourceFormat,
+      endpointPath,
+      providerSpecificData: credentials?.providerSpecificData,
+    });
  const responsesInputItems = Array.isArray(body?.input) ? body.input : [];
  const customToolNames = collectCustomToolNamesForSourceFormat(
    sourceFormat,
@@ -3387,6 +3392,15 @@ export async function handleChatCore({
        const responseHeaders = new Headers(headersObj);
        stripStaleForwardingHeaders(responseHeaders);
        stripNextMiddlewareControlHeaders(responseHeaders);
+        // The upstream headers (turn-state included) are about to be committed
+        // to the client — record which connection minted the blob so a later
+        // cross-account echo can be stripped (Codex failover guard).
+        if (provider === "codex" && readCodexTurnStateHeader(responseHeaders)) {
+          noteCodexTurnStateProvenance(
+            getCodexClientSessionId(clientRawRequest?.headers),
+            rawResult._executionCredentials?.connectionId ?? credentials?.connectionId
+          );
+        }
        const contentType = (responseHeaders.get("content-type") || "").toLowerCase();
        const payload = await readNonStreamingResponseBody(
          rawResult.response,
@@ -5106,6 +5120,17 @@ export async function handleChatCore({
    comboStrategy,
  });

+  // The streaming headers (turn-state included, when present) are committed to
+  // the client from here on — record which connection minted the blob so a
+  // later cross-account echo can be stripped (Codex failover guard). The
+  // in-place failover update means `credentials` is the winning account.
+  if (provider === "codex" && readCodexTurnStateHeader(providerResponse.headers)) {
+    noteCodexTurnStateProvenance(
+      getCodexClientSessionId(clientRawRequest?.headers),
+      credentials?.connectionId
+    );
+  }
+
  // Create transform stream with logger for streaming response
  let transformStream;
  const responseToolNameMap = mergeResponseToolNameMap(
--- a/open-sse/handlers/chatCore/responseHeaders.ts
+++ b/open-sse/handlers/chatCore/responseHeaders.ts
@@ -28,11 +28,18 @@ const STREAMING_RESPONSE_HEADER_DENYLIST = new Set([
  "x-amz-security-token",
  "x-auth-token",
  "x-accel-buffering",
-  // 314-byte Codex session blob. It is not a client rate-limit signal and
-  // alone ate ~40% of the old 768-byte budget, evicting x-codex-*-used-percent.
-  "x-codex-turn-state",
 ]);

+/**
+ * `x-codex-turn-state` is forwarded verbatim and EXEMPT from the forwarding
+ * budget. The real Codex client captures this ~314-byte blob from /responses
+ * (and echoes it back within the same turn), so dropping it breaks the
+ * protocol chain — but naively counting it against the budget used to evict
+ * the x-codex-*-used-percent quota headers (the reason it was denylisted
+ * under #10315-era budgeting). Carving it out keeps both.
+ */
+const CODEX_TURN_STATE_RESPONSE_HEADER = "x-codex-turn-state";
+
 const DEFAULT_FORWARDED_HEADER_BUDGET_BYTES = 768;

 /**
@@ -206,7 +213,9 @@ export function buildStreamingResponseHeaders(
      STREAMING_RESPONSE_HEADER_DENYLIST.has(normalized) ||
      connectionScopedHeaders.has(normalized) ||
      isNextMiddlewareControlHeader(normalized) ||
-      isOmniRouteInternalHeader(normalized)
+      isOmniRouteInternalHeader(normalized) ||
+      // Forwarded separately below, outside the byte budget.
+      normalized === CODEX_TURN_STATE_RESPONSE_HEADER
    ) {
      return;
    }
@@ -269,6 +278,10 @@ export function buildStreamingResponseHeaders(
    "X-Accel-Buffering": "no",
    [OMNIROUTE_RESPONSE_HEADERS.cache]: "MISS",
  };
+  const codexTurnState = providerHeaders.get(CODEX_TURN_STATE_RESPONSE_HEADER)?.trim();
+  if (codexTurnState) {
+    responseHeaders[CODEX_TURN_STATE_RESPONSE_HEADER] = codexTurnState;
+  }
  attachOmniRouteMetaHeaders(responseHeaders, meta);
  return responseHeaders;
 }
--- a/open-sse/handlers/cursorCliProxy.ts
+++ b/open-sse/handlers/cursorCliProxy.ts
@@ -0,0 +1,524 @@
+/**
+ * Cursor CLI passthrough.
+ *
+ * cursor-agent honours `--endpoint` / CURSOR_API_ENDPOINT and, with
+ * `network.useHttp1ForAgent: true`, talks to that endpoint exclusively over
+ * HTTP/1.1: unary Connect-RPC POSTs (`/aiserver.v1.*`, `/agent.v1.*`,
+ * `/aiserver.v1.BidiService/BidiAppend`), the agent turn as
+ * `/agent.v1.AgentService/RunSSE` (text/event-stream), OTLP traces on
+ * `/v1/traces`, and the API-key bootstrap `POST /auth/exchange_user_api_key`.
+ *
+ * Pointing the CLI at OmniRoute therefore only needs a thin forwarder:
+ *   1. `/auth/exchange_user_api_key` authenticates the CLI with an OmniRoute
+ *      API key and hands back an OmniRoute-minted session JWT. The CLI reads
+ *      `exp` from whatever JWT it receives and re-exchanges when the token is
+ *      opaque or expired, so the minted token must be a real JWT with `exp`.
+ *   2. Every other path verifies that JWT, resolves an active `cursor-api`
+ *      connection (the crsr_ key is exchanged for a session token), swaps the
+ *      Authorization header and streams the upstream reply back unchanged.
+ *      Each hop is recorded in call_logs.
+ */
+
+import { SignJWT, jwtVerify, type JWTPayload } from "jose";
+import { z } from "zod";
+import { getApiKeyById, getApiKeyMetadata, validateApiKey } from "@/lib/db/apiKeys";
+import { getProviderConnections } from "@/lib/db/providers";
+import { saveCallLog } from "@/lib/usage/callLogs";
+import { isRequireApiKeyEnabled } from "@/shared/utils/featureFlags";
+import { HTTP_STATUS } from "../config/constants.ts";
+import {
+  CURSOR_API_BASE_URL,
+  CURSOR_API_KEY_EXCHANGE_PATH,
+  CursorApiKeyExchangeError,
+  invalidateCursorSessionToken,
+  isCursorApiKey,
+  resolveCursorBearerToken,
+} from "../services/cursorApiKeyAuth.ts";
+import { sanitizeErrorMessage } from "../utils/error.ts";
+
+export const CURSOR_CLI_PROXY_PREFIX = "/api/cursor-cli";
+export const CURSOR_CLI_SESSION_ISSUER = "omniroute";
+export const CURSOR_CLI_SESSION_AUDIENCE = "cursor-cli";
+export const CURSOR_CLI_SESSION_TTL_SECONDS = 60 * 60;
+export const CURSOR_CLI_REQUEST_TYPE = "cursor-cli";
+const ANONYMOUS_SUBJECT = "anonymous";
+const PROVIDER_ID = "cursor-api";
+
+const REQUEST_HEADER_DENYLIST = new Set([
+  "authorization",
+  "host",
+  "connection",
+  "content-length",
+  "accept-encoding",
+  "keep-alive",
+  "proxy-authorization",
+  "proxy-connection",
+  "te",
+  "trailer",
+  "transfer-encoding",
+  "upgrade",
+  "x-forwarded-for",
+  "x-forwarded-host",
+  "x-forwarded-proto",
+  "x-real-ip",
+  "cookie",
+]);
+
+const RESPONSE_HEADER_DENYLIST = new Set([
+  "connection",
+  "content-encoding",
+  "content-length",
+  "keep-alive",
+  "transfer-encoding",
+  "set-cookie",
+]);
+
+const exchangeBodySchema = z.object({}).passthrough();
+
+const sessionClaimsSchema = z.object({
+  sub: z.string().min(1),
+  iss: z.literal(CURSOR_CLI_SESSION_ISSUER),
+  aud: z.union([
+    z.literal(CURSOR_CLI_SESSION_AUDIENCE),
+    z.array(z.string()).refine((list) => list.includes(CURSOR_CLI_SESSION_AUDIENCE)),
+  ]),
+  exp: z.number(),
+  name: z.string().nullable().optional(),
+});
+
+export type CursorCliPrincipal = {
+  apiKeyId: string | null;
+  apiKeyName: string | null;
+};
+
+export type CursorCliConnectionLike = {
+  id?: unknown;
+  apiKey?: unknown;
+  accessToken?: unknown;
+  priority?: unknown;
+  rateLimitedUntil?: unknown;
+};
+
+export type CursorCliProxyDeps = {
+  fetchImpl: typeof fetch;
+  now: () => number;
+  getSecret: () => string | undefined;
+  validateApiKey: (key: string) => Promise<boolean>;
+  getApiKeyMetadata: (key: string) => Promise<{ id: string; name: string } | null>;
+  getApiKeyById: (id: string) => Promise<{ isActive?: unknown; revokedAt?: unknown } | null>;
+  requireApiKey: () => boolean;
+  listCursorConnections: () => Promise<CursorCliConnectionLike[]>;
+  resolveBearer: (credentials: {
+    apiKey?: string | null;
+    accessToken?: string | null;
+  }) => Promise<string>;
+  invalidateBearer: (apiKey: string) => void;
+  saveCallLog: (entry: Record<string, unknown>) => Promise<void>;
+  upstreamBaseUrl: string;
+};
+
+const defaultDeps: CursorCliProxyDeps = {
+  fetchImpl: (input, init) => fetch(input, init),
+  now: () => Date.now(),
+  getSecret: () => process.env.JWT_SECRET,
+  validateApiKey: (key) => validateApiKey(key),
+  getApiKeyMetadata: async (key) => {
+    const meta = await getApiKeyMetadata(key);
+    return meta ? { id: meta.id, name: meta.name } : null;
+  },
+  getApiKeyById: (id) => getApiKeyById(id),
+  requireApiKey: () => isRequireApiKeyEnabled(),
+  listCursorConnections: async () =>
+    (await getProviderConnections({
+      provider: PROVIDER_ID,
+      isActive: true,
+    })) as CursorCliConnectionLike[],
+  resolveBearer: (credentials) => resolveCursorBearerToken(credentials),
+  invalidateBearer: (apiKey) => invalidateCursorSessionToken(apiKey),
+  saveCallLog: (entry) => saveCallLog(entry),
+  upstreamBaseUrl: CURSOR_API_BASE_URL,
+};
+
+function jsonResponse(status: number, body: Record<string, unknown>): Response {
+  return new Response(JSON.stringify(body), {
+    status,
+    headers: { "content-type": "application/json" },
+  });
+}
+
+function connectError(status: number, code: string, message: string): Response {
+  return jsonResponse(status, { code, message: sanitizeErrorMessage(message) });
+}
+
+function extractBearer(request: Request): string | null {
+  const header = request.headers.get("authorization") ?? "";
+  const match = /^Bearer\s+(.+)$/i.exec(header.trim());
+  return match ? match[1].trim() : null;
+}
+
+function secretKey(secret: string): Uint8Array {
+  return new TextEncoder().encode(secret);
+}
+
+export function normalizeCursorCliPath(segments: readonly string[]): string {
+  return "/" + segments.map((segment) => encodeURIComponent(decodeURIComponent(segment))).join("/");
+}
+
+async function authenticateExchange(
+  request: Request,
+  deps: CursorCliProxyDeps
+): Promise<CursorCliPrincipal | Response> {
+  const bearer = extractBearer(request);
+  if (bearer && (await deps.validateApiKey(bearer))) {
+    const meta = await deps.getApiKeyMetadata(bearer);
+    return { apiKeyId: meta?.id ?? null, apiKeyName: meta?.name ?? null };
+  }
+  if (!deps.requireApiKey()) {
+    return { apiKeyId: null, apiKeyName: null };
+  }
+  return connectError(
+    HTTP_STATUS.UNAUTHORIZED,
+    "unauthenticated",
+    "CURSOR_API_KEY must be an OmniRoute API key when OmniRoute requires API keys"
+  );
+}
+
+export async function mintCursorCliSessionToken(
+  principal: CursorCliPrincipal,
+  secret: string,
+  nowMs: number
+): Promise<string> {
+  const nowSeconds = Math.floor(nowMs / 1000);
+  return new SignJWT({ name: principal.apiKeyName })
+    .setProtectedHeader({ alg: "HS256", typ: "JWT" })
+    .setIssuer(CURSOR_CLI_SESSION_ISSUER)
+    .setAudience(CURSOR_CLI_SESSION_AUDIENCE)
+    .setSubject(principal.apiKeyId ?? ANONYMOUS_SUBJECT)
+    .setIssuedAt(nowSeconds)
+    .setExpirationTime(nowSeconds + CURSOR_CLI_SESSION_TTL_SECONDS)
+    .sign(secretKey(secret));
+}
+
+async function verifyCursorCliSessionToken(
+  token: string,
+  secret: string,
+  nowMs: number
+): Promise<CursorCliPrincipal | null> {
+  let payload: JWTPayload;
+  try {
+    ({ payload } = await jwtVerify(token, secretKey(secret), {
+      issuer: CURSOR_CLI_SESSION_ISSUER,
+      audience: CURSOR_CLI_SESSION_AUDIENCE,
+      currentDate: new Date(nowMs),
+    }));
+  } catch {
+    return null;
+  }
+  const claims = sessionClaimsSchema.safeParse(payload);
+  if (!claims.success) return null;
+  return {
+    apiKeyId: claims.data.sub === ANONYMOUS_SUBJECT ? null : claims.data.sub,
+    apiKeyName: claims.data.name ?? null,
+  };
+}
+
+async function isPrincipalStillValid(
+  principal: CursorCliPrincipal,
+  deps: CursorCliProxyDeps
+): Promise<boolean> {
+  if (!principal.apiKeyId) return !deps.requireApiKey();
+  const row = await deps.getApiKeyById(principal.apiKeyId);
+  if (!row) return false;
+  if (row.isActive === false) return false;
+  return !(typeof row.revokedAt === "string" && row.revokedAt.trim() !== "");
+}
+
+type ResolvedConnection = {
+  connectionId: string | null;
+  bearer: string;
+  apiKey: string | null;
+};
+
+function connectionPriority(connection: CursorCliConnectionLike): number {
+  return typeof connection.priority === "number" ? connection.priority : Number.MAX_SAFE_INTEGER;
+}
+
+function isCoolingDown(connection: CursorCliConnectionLike, nowMs: number): boolean {
+  if (typeof connection.rateLimitedUntil !== "string") return false;
+  const until = Date.parse(connection.rateLimitedUntil);
+  return Number.isFinite(until) && until > nowMs;
+}
+
+async function resolveUpstreamConnection(
+  deps: CursorCliProxyDeps
+): Promise<ResolvedConnection | Response> {
+  const connections = (await deps.listCursorConnections())
+    .filter((connection) => !isCoolingDown(connection, deps.now()))
+    .sort((a, b) => connectionPriority(a) - connectionPriority(b));
+  if (connections.length === 0) {
+    return connectError(
+      HTTP_STATUS.SERVICE_UNAVAILABLE,
+      "unavailable",
+      "No active Cursor API connection configured in OmniRoute"
+    );
+  }
+  let lastError: unknown = null;
+  for (const connection of connections) {
+    const apiKey = isCursorApiKey(connection.apiKey) ? connection.apiKey : null;
+    const accessToken = typeof connection.accessToken === "string" ? connection.accessToken : null;
+    try {
+      const bearer = await deps.resolveBearer({ apiKey, accessToken });
+      return {
+        connectionId: typeof connection.id === "string" ? connection.id : null,
+        bearer,
+        apiKey,
+      };
+    } catch (err) {
+      lastError = err;
+    }
+  }
+  const status =
+    lastError instanceof CursorApiKeyExchangeError ? lastError.status : HTTP_STATUS.BAD_GATEWAY;
+  const message = lastError instanceof Error ? lastError.message : "Cursor credential unavailable";
+  return connectError(
+    status,
+    status === HTTP_STATUS.UNAUTHORIZED ? "unauthenticated" : "unavailable",
+    message
+  );
+}
+
+function buildUpstreamHeaders(request: Request, bearer: string): Headers {
+  const headers = new Headers();
+  request.headers.forEach((value, name) => {
+    if (!REQUEST_HEADER_DENYLIST.has(name.toLowerCase())) headers.set(name, value);
+  });
+  headers.set("authorization", `Bearer ${bearer}`);
+  return headers;
+}
+
+function buildDownstreamHeaders(upstream: Response): Headers {
+  const headers = new Headers();
+  upstream.headers.forEach((value, name) => {
+    if (!RESPONSE_HEADER_DENYLIST.has(name.toLowerCase())) headers.set(name, value);
+  });
+  return headers;
+}
+
+type CallLogInput = {
+  method: string;
+  path: string;
+  status: number;
+  startedAt: number;
+  principal: CursorCliPrincipal | null;
+  connectionId: string | null;
+  error?: string | null;
+};
+
+function recordCall(deps: CursorCliProxyDeps, input: CallLogInput): void {
+  void deps
+    .saveCallLog({
+      method: input.method,
+      path: `${CURSOR_CLI_PROXY_PREFIX}${input.path}`,
+      status: input.status,
+      model: "-",
+      provider: PROVIDER_ID,
+      connectionId: input.connectionId,
+      duration: Math.max(0, deps.now() - input.startedAt),
+      apiKeyId: input.principal?.apiKeyId ?? null,
+      apiKeyName: input.principal?.apiKeyName ?? null,
+      requestType: CURSOR_CLI_REQUEST_TYPE,
+      sourceFormat: CURSOR_CLI_REQUEST_TYPE,
+      targetFormat: CURSOR_CLI_REQUEST_TYPE,
+      error: input.error ? { message: sanitizeErrorMessage(input.error) } : null,
+    })
+    .catch(() => undefined);
+}
+
+function streamWithCompletionLog(
+  body: ReadableStream<Uint8Array>,
+  onDone: (error?: string) => void
+): ReadableStream<Uint8Array> {
+  const reader = body.getReader();
+  let settled = false;
+  const settle = (error?: string) => {
+    if (settled) return;
+    settled = true;
+    onDone(error);
+  };
+  return new ReadableStream<Uint8Array>({
+    async pull(controller) {
+      try {
+        const { done, value } = await reader.read();
+        if (done) {
+          settle();
+          controller.close();
+          return;
+        }
+        controller.enqueue(value);
+      } catch (err) {
+        settle(err instanceof Error ? err.message : "upstream stream failed");
+        controller.error(err);
+      }
+    },
+    cancel(reason) {
+      settle(reason instanceof Error ? reason.message : "stream cancelled");
+      return reader.cancel(reason);
+    },
+  });
+}
+
+async function handleExchange(
+  request: Request,
+  startedAt: number,
+  deps: CursorCliProxyDeps
+): Promise<Response> {
+  if (request.method !== "POST") {
+    return connectError(405, "unimplemented", "Use POST");
+  }
+  const rawBody = await request.text();
+  if (rawBody.trim().length > 0) {
+    let parsed: unknown;
+    try {
+      parsed = JSON.parse(rawBody);
+    } catch {
+      return connectError(HTTP_STATUS.BAD_REQUEST, "invalid_argument", "Body must be JSON");
+    }
+    if (!exchangeBodySchema.safeParse(parsed).success) {
+      return connectError(
+        HTTP_STATUS.BAD_REQUEST,
+        "invalid_argument",
+        "Body must be a JSON object"
+      );
+    }
+  }
+
+  const principal = await authenticateExchange(request, deps);
+  if (principal instanceof Response) {
+    recordCall(deps, {
+      method: request.method,
+      path: CURSOR_API_KEY_EXCHANGE_PATH,
+      status: principal.status,
+      startedAt,
+      principal: null,
+      connectionId: null,
+      error: "OmniRoute API key rejected",
+    });
+    return principal;
+  }
+
+  const secret = deps.getSecret();
+  if (!secret || secret.trim().length === 0) {
+    return connectError(
+      HTTP_STATUS.SERVICE_UNAVAILABLE,
+      "unavailable",
+      "JWT_SECRET is not configured; the Cursor CLI passthrough cannot mint session tokens"
+    );
+  }
+
+  const token = await mintCursorCliSessionToken(principal, secret, deps.now());
+  recordCall(deps, {
+    method: request.method,
+    path: CURSOR_API_KEY_EXCHANGE_PATH,
+    status: 200,
+    startedAt,
+    principal,
+    connectionId: null,
+  });
+  return jsonResponse(200, { accessToken: token, refreshToken: token });
+}
+
+async function handleForward(
+  request: Request,
+  path: string,
+  startedAt: number,
+  deps: CursorCliProxyDeps
+): Promise<Response> {
+  const secret = deps.getSecret();
+  const bearer = extractBearer(request);
+  const principal =
+    bearer && secret ? await verifyCursorCliSessionToken(bearer, secret, deps.now()) : null;
+  if (!principal || !(await isPrincipalStillValid(principal, deps))) {
+    return connectError(
+      HTTP_STATUS.UNAUTHORIZED,
+      "unauthenticated",
+      "Missing or expired OmniRoute Cursor CLI session token"
+    );
+  }
+
+  const resolved = await resolveUpstreamConnection(deps);
+  if (resolved instanceof Response) {
+    recordCall(deps, {
+      method: request.method,
+      path,
+      status: resolved.status,
+      startedAt,
+      principal,
+      connectionId: null,
+      error: "No usable Cursor connection",
+    });
+    return resolved;
+  }
+
+  const search = new URL(request.url).search;
+  const upstreamUrl = `${deps.upstreamBaseUrl}${path}${search}`;
+  const hasBody = request.method !== "GET" && request.method !== "HEAD";
+  let upstream: Response;
+  try {
+    upstream = await deps.fetchImpl(upstreamUrl, {
+      method: request.method,
+      headers: buildUpstreamHeaders(request, resolved.bearer),
+      body: hasBody ? request.body : undefined,
+      signal: request.signal,
+      redirect: "manual",
+      ...(hasBody ? { duplex: "half" } : {}),
+    } as RequestInit);
+  } catch (err) {
+    const message = err instanceof Error ? err.message : "upstream request failed";
+    recordCall(deps, {
+      method: request.method,
+      path,
+      status: HTTP_STATUS.BAD_GATEWAY,
+      startedAt,
+      principal,
+      connectionId: resolved.connectionId,
+      error: message,
+    });
+    return connectError(HTTP_STATUS.BAD_GATEWAY, "unavailable", message);
+  }
+
+  if (upstream.status === HTTP_STATUS.UNAUTHORIZED && resolved.apiKey) {
+    deps.invalidateBearer(resolved.apiKey);
+  }
+
+  const logInput: CallLogInput = {
+    method: request.method,
+    path,
+    status: upstream.status,
+    startedAt,
+    principal,
+    connectionId: resolved.connectionId,
+  };
+  const headers = buildDownstreamHeaders(upstream);
+  if (!upstream.body) {
+    recordCall(deps, logInput);
+    return new Response(null, { status: upstream.status, headers });
+  }
+  const body = streamWithCompletionLog(upstream.body, (error) =>
+    recordCall(deps, { ...logInput, error: error ?? null })
+  );
+  return new Response(body, { status: upstream.status, headers });
+}
+
+export async function handleCursorCliProxy(
+  request: Request,
+  segments: readonly string[],
+  overrides: Partial<CursorCliProxyDeps> = {}
+): Promise<Response> {
+  const deps: CursorCliProxyDeps = { ...defaultDeps, ...overrides };
+  const startedAt = deps.now();
+  const path = normalizeCursorCliPath(segments);
+  if (path === CURSOR_API_KEY_EXCHANGE_PATH) {
+    return handleExchange(request, startedAt, deps);
+  }
+  return handleForward(request, path, startedAt, deps);
+}
--- a/open-sse/mcp-server/httpTransport.ts
+++ b/open-sse/mcp-server/httpTransport.ts
@@ -284,12 +284,30 @@ export async function handleMcpStreamableHTTP(request: Request): Promise<Respons
  return protectMcpSseResponse(request, await handleStreamableRequest(request));
 }

+interface RpcRequest {
+  method?: string;
+  [key: string]: unknown;
+}
+
 /**
 * Handle SSE requests.
 * SSE transport is implemented via Streamable HTTP transport with GET for SSE stream
 * and POST for messages (the Streamable HTTP transport supports both patterns).
 */
 export async function handleMcpSSE(request: Request): Promise<Response> {
+  if (request.method === "POST") {
+    try {
+      const body = await request.clone().json();
+      const isInitialize = Array.isArray(body)
+        ? body.some((req: RpcRequest) => req?.method === "initialize")
+        : (body as RpcRequest)?.method === "initialize";
+
+      if (isInitialize) {
+        console.log("[MCP] New client initialize detected, resetting SSE singleton...");
+        closeSseTransport();
+      }
+    } catch (err) {}
+  }
  const { transport } = ensureSseServer();

  try {
--- a/open-sse/services/codexQuotaFetcher.ts
+++ b/open-sse/services/codexQuotaFetcher.ts
@@ -25,6 +25,7 @@ import {
 import { registerQuotaFetcher, registerQuotaWindows, type QuotaInfo } from "./quotaPreflight.ts";
 import { registerMonitorFetcher } from "./quotaMonitor.ts";
 import { throttleQuotaFetch } from "./quotaFetchThrottle.ts";
+import { getCodexBackendIdentityHeaders } from "../config/codexClient.ts";

 /**
 * Stable identifiers for Codex's quota windows. These match the quota keys
@@ -222,6 +223,9 @@ export async function fetchCodexQuota(
      Authorization: `Bearer ${meta.accessToken}`,
      "Content-Type": "application/json",
      Accept: "application/json",
+      // Canonical Codex backend identity (UA + originator + version), same
+      // chain as inference — see getCodexUsage.
+      ...getCodexBackendIdentityHeaders(),
    };

    if (meta.workspaceId) {
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Diego Rodrigues de Sa e Souza	0f13fe4221	feat(radar): stable catalog export workflow with provenance (#10826 ) Passo 10 of the Radar go-live: publish the OmniRoute catalog export the Radar server consumes to a stable URL, so the 1 GB-RAM private server pulls it via RADAR_EXPORT_URL instead of depending on the deploy-time snapshot. - scripts/release/radar-export.mjs: emits {geradoEm, budgets, totais, registry, provenance} from the catalog config modules. Provenance (sourceCommit, sourceRef, runUrl, generatedBy) is never fabricated — unknown fields stay null. - .github/workflows/radar-export.yml: on main catalog changes / manual dispatch / weekly, generates the export and clobbers the stable 'radar-export-latest' release asset (gh release, GH_TOKEN — checkout persist-credentials:false). - tests/unit/radar-export.test.mjs: consumer contract (budgets[] non-empty) + provenance null-when-unknown + GitHub-env reflection. Stable URL for RADAR_EXPORT_URL: https://github.com/diegosouzapw/OmniRoute/releases/download/radar-export-latest/export-omniroute.json Re-baselines zizmorFindings 190->192 (+2 unpinned-uses @vN, the repo-wide deliberate convention; artipacked auto-fixed). Co-authored-by: Xiangzhe <bakryun0718@proton.me>	2026-08-20 08:09:42 -03:00
Markus Hartung	5089c17b44	fix(resilience): scope same-account transport retry out of emergency-fallback and combo hops #10792 (#9708) added a same-account retry for retryable 502/503/504/507 transport failures, applied uniformly inside handleSingleModelChat. Two other paths call into the same function recursively/iteratively and each carries its own documented single-call guarantee that the retry silently broke: - Emergency fallback (#1731): exactly one hop to the free fallback model, no extra calls against an already-exhausted provider. The retry was doubling that call whenever the fallback model itself returned a transient-looking status. - Combo routing: target-level fallback is the combo's own policy (next target, not same-account retry). The retry delayed that policy and could surface the wrong terminal status when a later combo/global-fallback hop threw. Both regressions were already covered by existing tests in chat-route-coverage.test.ts (asserting exact call counts / preserved status) — confirmed red on the release tip before this fix, green after. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>	2026-08-20 08:08:09 -03:00
Diego Rodrigues de Sa e Souza	bc6129bcb2	fix(relay): normalize bifrost errors, remap credential 404, fix analytics (#10797 ) Merged — the 5 pre-existing tests that broke from this PR's intentional 404→401 remap (single-model no-credentials) are now realigned to the new contract. Thanks!	2026-08-20 06:37:02 -03:00
Markus Hartung	3810a6c52c	fix(tests): realign provider count to 230 after merge-train batch1 collision #10722 (Token Kiosk) and #10729 (Cursor) each independently bumped the gateways/specialty-media APIKEY_PROVIDERS count by one in the same merge-train batch; neither PR could see the other's bump, so the merged tree landed at 230 while the surviving hardcoded assertion still said 229. Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>	2026-08-20 06:31:49 -03:00
Michael YC JO	85aa84ebcb	fix(tests): realign two more stream-utils passthrough cases with the #10017 control-line rules (#10796 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:46 -03:00
Prudhvi Vuda	c00086e616	fix(catalog): exclude generic 128k default from combo context min() (#10734 ) (#10793 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:43 -03:00
Prudhvi Vuda	aa32d2ed77	fix(resilience): retry Codex pre-output transport failures on the same account (#9708 ) (#10792 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:39 -03:00
Prudhvi Vuda	71f858fc48	fix(translator): convert OpenAI tool-result images to Claude image blocks (#9692 ) (#10791 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:36 -03:00
Dizzle	6307504584	feat(routing): make task-detection patterns operator-configurable (#10783 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:31 -03:00
Dizzle	404d3a4211	fix(sse): add a protocol-level ping/pong heartbeat to LiveWS (#10782 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:20 -03:00
Dizzle	d82e3cf8e9	fix(db): periodically truncate the SQLite WAL on long-running servers (#10781 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:16 -03:00
Dizzle	121023e418	docs: document combo invocation by exact model name (#7992 ) (#10779 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:12 -03:00
Farshid Rezaei	4fc0b412fe	add fa readme (#10777 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:08 -03:00
Dizzle	66144d86f2	chore(security): drop the unused enforceSecrets() duplicate (#10775 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:05 -03:00
Jorge Delgado	f7cba50cb7	fix(analytics): treat Claude Code (`claude`/`cc`) as a flat-rate subscription provider (#10773 ) (#10774 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:30:01 -03:00
Sahil Singh	6ff83077fd	fix(api): reset mcp sse singleton on new client initialize (#10690 ) (#10772 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:49 -03:00
Dizzle	49a47cbe6f	feat(api): answer GET /api/health without a key (#10771 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:45 -03:00
Dizzle	ee230fa93a	fix(logging): keep the component and apply printf formats in the app log (#10770 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:42 -03:00
Dizzle	7d2efdf1a4	fix(api): report the cache that requests actually use in /api/cache/stats (#10769 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:38 -03:00
InkshadeWoods	5100642ebb	fix(sensenova): clamp max reasoning effort to xhigh (#10733 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:34 -03:00
acc0mplish	2eb6d59ebc	fix(providers): copilot-m365-web invocation dropped by substrate + token refresh (#10732 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:30 -03:00
pandaaaa1990	48f3428307	pUpdate python_requests.py (#10731 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:18 -03:00
stanley	bc9a685b70	feat(combo): opaque per-invocation decision trace for priority fallbacks (#10681 ) (#10730 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:14 -03:00
Damian Pozimski	3e0afc8444	feat(providers): Cursor API provider (crsr_ keys) and Cursor CLI passthrough (#10729 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:11 -03:00
Michael YC JO	83c77fb0bf	fix(oauth): keep the Kiro profileArn on IAM Identity Center logins (#10725 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:08 -03:00
hgaib	08f23d0d0d	feat(providers): add Token Kiosk as an OpenAI-compatible provider (#10722 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:29:03 -03:00
echoriver89	5185571f18	fix(electron): pin embedded server to loopback so the window shows on Windows (#10717 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:59 -03:00
Xiangzhe	f330b21afd	feat(codex): sync v178 identity mechanisms — turn-state relay, persisted seeds, identity faces (#10716 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:45 -03:00
Nguyen Thanh Dat	b11b000048	fix(agent-bridge): make the regenerate-cert endpoint actually mint a new cert (#10467 ) (#10715 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:41 -03:00
Xiangzhe	82e5afed6b	feat(usage): show Kimi Coding Extra Usage (#10712 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:37 -03:00
Webman	1accabeb4e	fix(db): prevent Windows native-driver hang from stalling requests (#10627 ) (#10709 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:33 -03:00
Ara	81b0ff46a3	fix(cline): label internal health checks (#10706 ) Merged via merge-train (release/v3.8.50, batch1 2026-08-20) — static gates (typecheck/file-size/complexity/cognitive/changelog) green on the combined tree; test:unit reds observed in the boarded run were verified pre-existing on the pure release tip (unrelated flake), not caused by this PR. Thanks for the contribution!	2026-08-20 06:28:29 -03:00
Markus Hartung	41ffb08e4a	chore(quality): rebaseline file-size for merge-train batch1 (#10722 , #10797 ) + pre-existing chatBodyAdmission.ts drift Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>	2026-08-20 00:50:35 -03:00
Diego Rodrigues de Sa e Souza	3d7ed7aa87	fix(build): tolerate same-realpath symlink / stale-typed dest in assembleStandalone (#10776 ) Under heavy concurrent build I/O, the bulk .build/next/standalone -> outDir tree copy can already have carried a prior pass's result into a NATIVE_ASSET_ENTRIES/EXTRA_MODULE_ENTRIES dest before that entry's own copy runs (an absolute pnpm-store symlink resolving to the exact same realpath as src, or a stale node of a different type). fs.cpSync/fs.cp refuse to overwrite either case even with force:true, throwing ERR_FS_CP_EINVAL ("src and dest cannot be the same") or ERR_FS_CP_DIR_TO_NON_DIR/ERR_FS_CP_NON_DIR_TO_DIR — non-deterministically crashing the build:release/build:cli deploy pipeline on whichever entry the race happened to hit that run. Adds resolvesToSamePath/clearStaleDest guards to all four copy call sites (the two sync loops in copyNativeAssetsAndExtraModules, repairEmptyExternalPackageDirs, and the async syncNativeAssetsToDir/syncExtraModulesToDir twins) so a dest already pointing at src is skipped and any other stale occupant is cleared before the fresh copy. Co-authored-by: Markus Hartung <mail@hartmark.se>	2026-08-19 17:58:58 -03:00
Markus Hartung	3c9cb21cca	feat(ad-hoc): add mesh interaction scripts for Discord integration	2026-08-19 15:33:25 -03:00
				`@@ -0,0 +1 @@`
				- feat(providers): new `cursor-api` provider (card "Cursor API", alias `cua`): connect a Cursor user API key (`crsr_…`) and route `cursor-api/<model>` through the existing Cursor agent executor (the key is exchanged for a 1h session token and cached), plus a `/api/cursor-cli/*` passthrough so the Cursor CLI itself runs through OmniRoute (`CURSOR_API_ENDPOINT=http://<omniroute>/api/cursor-cli`, `CURSOR_API_KEY=<OmniRoute key>`) with every RPC attributed and logged. The IDE `cursor` provider is unchanged. (#10729)
				`@@ -0,0 +1 @@`
				- feat(api): `GET /api/health` now answers `{ status, timestamp }` without a key. Until now the path had no route, so the management-auth boundary answered first with a 401 — indistinguishable from a wrong key or an unknown route, which left Docker HEALTHCHECKs and Kubernetes probes unable to tell "down" from "misconfigured". Kept deliberately minimal: version, uptime and memory stay behind the authenticated `/api/monitoring/health` ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10771)).
				`@@ -0,0 +1 @@`
				- feat(routing): make Task-Aware Smart Routing's detection patterns operator-configurable via `settings.taskRouting.patternOverrides` (`PUT /api/settings/task-routing`) — the built-in patterns are English-only, so a non-English dashboard had no recourse short of turning detection off entirely; an override now replaces the pattern list for one task type without touching the rest (#10783)
				`@@ -0,0 +1 @@`
				`- feat(usage): show Kimi Coding's fixed-order Code 5-hour/7-day quota windows plus Extra Usage status, balance, monthly spend/limit, and the official Additional Credits link on Dashboard → Quota cards.`
				`@@ -0,0 +1 @@`
				- fix(providers): copilot-m365-web chat turns no longer surface as `(empty response)` — the type:4 invocation is aligned with the 2026-08 wire shape and now carries its type:1 Metrics follow-up in the same socket write, and the access token pre-flight-refreshes from a stored refresh_token instead of requiring a DevTools re-capture every ~75 minutes ([#10732](https://github.com/diegosouzapw/OmniRoute/pull/10732) — thanks @acc0mplish)
				`@@ -0,0 +1 @@`
				- fix(catalog): stop counting `getTokenLimit()`'s generic 128k catch-all as a known combo window, so `/v1/models` advertises the min of sourced member contexts instead of collapsing a 500k combo to 128k ([#10734](https://github.com/diegosouzapw/OmniRoute/issues/10734))
				`@@ -0,0 +1 @@`
				- fix(api): `/api/cache/stats` reported the prompt-cache LRU, which no request path ever writes to — it answered `0 hit / 0 miss, size 0` while the semantic cache served real traffic, and the Health and Usage dashboards rendered that as fact. It now reports the semantic cache's in-memory entries, with the same response shape ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10769)) — thanks @Poid-ZA, who first fixed this in #9446.
				`@@ -0,0 +1 @@`
				- fix(logging): the app log is filterable and readable again. Entries from the tagged logger (`[LEVEL] [TAG] message`) were filed under the level instead of the component, and printf format strings were never applied, so `%s`/`%d` stayed literal with the values trailing behind them unlabelled — including every LiveWS connection line, where the format is deliberate hardening against injected format specifiers ([#PRNUM](https://github.com/diegosouzapw/OmniRoute/pull/10770)).
				`@@ -0,0 +1 @@`
				- fix(analytics): Claude Code (`claude`/`cc`) is a flat-rate subscription, so cost analytics reports `$0` for it instead of estimating Anthropic list prices — the metered `anthropic` API keeps its real cost, and budget/quota/routing still estimate as before ([#10774](https://github.com/diegosouzapw/OmniRoute/pull/10774)) — thanks @electrumguy