diegosouzapw
b08fb31a28
feat(gateway): Phase 9 — LLM Gateway Intelligence
...
9.1 — Semantic Cache
- New: src/lib/semanticCache.js — Two-tier cache (in-memory LRU + SQLite)
- Signature = SHA-256(model + normalized messages + temperature + top_p)
- Only caches non-streaming, temperature=0 requests
- X-OmniRoute-No-Cache: true header bypass
- Response headers: X-OmniRoute-Cache: HIT/MISS
- DB table: semantic_cache with indexes on signature and model
- New: src/app/api/cache/route.js — GET stats, DELETE clear
9.2 — Request Idempotency
- New: src/lib/idempotencyLayer.js — In-memory 5s dedup window
- Reads Idempotency-Key or X-Request-Id headers
- Returns cached response with X-OmniRoute-Idempotent: true header
- Ephemeral by design (no SQLite)
9.3 — Progress Tracking in Streaming
- New: open-sse/utils/progressTracker.js
- Emits SSE 'event: progress' with tokens_generated + elapsed_ms
- Opt-in via X-OmniRoute-Progress: true header
- Supports AbortSignal cancellation
- Final event includes done: true
Integration:
- chatCore.js: idempotency check → cache check → provider call → cache store → idempotency save
- Streaming path: optional progress transform chain
- DB: semantic_cache table added to db/core.js schema
Tests: 320 pass (+25 new) | Build: success
2026-02-15 11:04:51 -03:00
diegosouzapw
6e29bd1197
feat(resilience): rate limit overhaul — exponential backoff, circuit breaker, anti-thundering herd, Resilience UI
...
Phase 1: Error classification + provider profiles (constants, providerRegistry, accountFallback)
Phase 2: Circuit breaker integration in combo pipeline (combo.js)
Phase 3: Anti-thundering herd mutex + auto rate limits for API key providers (auth.js, rateLimitManager)
Phase 4: Frontend Resilience tab with Circuit Breaker, Provider Profiles, Rate Limit cards
Tests: 63/63 passing (error-classification, combo-circuit-breaker, thundering-herd, rate-limit-enhanced)
2026-02-15 01:42:50 -03:00
diegosouzapw
5db6676319
feat: detect and handle unrecoverable refresh token errors by marking connections as expired and requiring re-authentication.
2026-02-14 22:04:00 -03:00
diegosouzapw
81a4f2986c
feat: v0.2.0 — advanced routing services, cost analytics dashboard, pricing overhaul
...
Added:
- 8 new open-sse services (account selector, IP filter, session manager, etc.)
- 6 new dashboard settings tabs (IP filter, system prompt, thinking budget, pricing)
- Usage cost dashboard with provider cost donut, cost trend line, model cost column
- Pricing API merging registry + custom + pricing-only models
- 9 unit tests for all new services
Changed:
- Usage analytics layout redesigned with prominent cost display
- DailyTrendChart upgraded to ComposedChart with dual Y-axes
Fixed:
- Pricing page now shows custom/imported models
- Icon rendering (material-symbols-rounded → outlined)
2026-02-14 13:50:45 -03:00
diegosouzapw
c6000ecc8e
feat(providers): add new endpoints (rerank, audio, moderations) and providers (Hyperbolic, Deepgram, AssemblyAI, NanoBanana)
...
- Add /v1/rerank endpoint with Cohere, Together, NVIDIA, Fireworks
- Add /v1/audio/transcriptions with OpenAI, Groq, Deepgram, AssemblyAI
- Add /v1/audio/speech with OpenAI, Hyperbolic, Deepgram
- Add /v1/moderations with OpenAI
- Add Hyperbolic as chat provider (8 models, OpenAI-compatible)
- Add Hyperbolic image generation (SDXL, SD2, FLUX)
- Add NanoBanana image generation (Flash + Pro via nanobananaapi.ai)
- Add Deepgram STT (Nova 3, Nova 2) with Token auth and binary format
- Add AssemblyAI STT (Universal 3 Pro) with async upload-poll workflow
- Add Deepgram TTS (Aura voices) and Hyperbolic TTS (Melo)
- Update EndpointPageClient to show 7 endpoint sections
- Update /v1/models to return type/subtype for all model categories
- Fix build: remove output:standalone from next.config.mjs
2026-02-13 22:57:32 -03:00
diegosouzapw
50fc931086
feat(versioning): set initial version to 0.0.1 and restore dynamic model import
...
- Set package.json version to 0.0.1 (initial OmniRoute release)
- Set open-sse/package.json version to 0.0.1
- Restore 'Import from /models' button for standard providers (openai, gemini, deepseek, etc.)
- Add handleImportModels function to ProviderDetailPage
2026-02-13 16:49:06 -03:00
diegosouzapw
699329944e
feat: initial OmniRoute release (rebranded from 9router)
...
This project is inspired by and originally forked from 9router by decolua
(https://github.com/decolua/9router ).
Full rebrand: 9router → OmniRoute across all source code, configuration,
Docker, documentation, and assets.
2026-02-13 16:29:27 -03:00
