feat(providers): add Pioneer AI (Fastino Labs) provider (#4909)

Integrated into release/v3.8.37 — providers.ts apikey block re-homed onto god-file-split src/shared/constants/providers/apikey/frontier-labs.ts (inline APIKEY_PROVIDERS no longer exists); registry/pioneer + providers/index.ts applied clean; 6/6 green.

CHANGELOG.md

@@ -10,6 +10,8 @@ _In development — bullets added per PR; finalized at release._
 
 ### 🔧 Bug Fixes
 
+- **feat(providers):** add Pioneer AI (Fastino Labs) provider — OpenAI-compatible chat completions at `api.pioneer.ai/v1`. Registered with alias `pn`, `X-API-Key` auth, and a catalog of 10 open-tier serverless models (Qwen3, Llama 3.1/3.2, Gemma 3, SmolLM3). Free $75 credits, no credit card required. Gated enterprise models (Claude/GPT/Gemini) require prior fine-tuning on the Pioneer platform and are intentionally excluded from the catalog. (thanks @HikiNarou)
+
 - **feat(sse): auto-promote successful combo model**: a new opt-in `comboAutoPromoteEnabled` setting reorders a combo's persisted model list so that, when a combo model responds successfully, it is moved to position #1 for future requests. (thanks @arssnndr)
 
 - **fix(sse):** dense, deterministic `response.output` ordering in `response.completed` — items are now sorted by their actual `output_index` (via a recorded-as-emitted accumulator + stable sort) instead of being rebuilt from unordered state dicts; `normalizeOutputIndex` replaces fragile `parseInt` calls for robust index coercion; superseded tool calls (replaced at the same index mid-stream) are excluded from the final output array. (thanks @Marco9113)

open-sse/config/providers/index.ts

@@ -167,6 +167,7 @@ import { openadapterProvider } from "./registry/openadapter/index.ts";
 import { ditProvider } from "./registry/dit/index.ts";
 import { tokenrouterProvider } from "./registry/tokenrouter/index.ts";
 import { codebuddy_cnProvider } from "./registry/codebuddy-cn/index.ts";
+import { pioneerProvider } from "./registry/pioneer/index.ts";
 
 export const REGISTRY: Record<string, RegistryEntry> = {
   aimlapi: aimlapiProvider,
@@ -336,4 +337,5 @@ export const REGISTRY: Record<string, RegistryEntry> = {
   dit: ditProvider,
   tokenrouter: tokenrouterProvider,
   "codebuddy-cn": codebuddy_cnProvider,
+  pioneer: pioneerProvider,
 };

open-sse/config/providers/registry/pioneer/index.ts

@@ -0,0 +1,41 @@
+import type { RegistryEntry } from "../../shared.ts";
+
+/**
+ * Pioneer AI by Fastino Labs — OpenAI-compatible chat completions.
+ *
+ * Endpoint: https://api.pioneer.ai/v1/chat/completions
+ * Auth: X-API-Key header with a pio_sk_... key (Bearer also accepted upstream).
+ *
+ * Only models with supports_on_demand_inference=true work directly with a bare
+ * pio_sk_ key. Gated/enterprise models (Claude/GPT/Gemini etc.) require a prior
+ * fine-tuning job and are called via the resulting job id, not the base model id.
+ *
+ * Source of truth for models: GET https://api.pioneer.ai/base-models
+ *   ?supports_inference=true&task_type=decoder
+ *   filter: supports_on_demand_inference=true
+ */
+export const pioneerProvider: RegistryEntry = {
+  id: "pioneer",
+  alias: "pn",
+  format: "openai",
+  executor: "default",
+  baseUrl: "https://api.pioneer.ai/v1/chat/completions",
+  authType: "apikey",
+  // Pioneer standardises on X-API-Key (preferred over Bearer).
+  // The default executor resolves "x-api-key" to the X-API-Key header.
+  authHeader: "x-api-key",
+  models: [
+    // === Open-tier serverless inference ===
+    // These models support on-demand inference with any pio_sk_ key.
+    { id: "Qwen/Qwen3-32B", name: "Qwen3 32B" },
+    { id: "Qwen/Qwen3.6-27B", name: "Qwen3.6 27B" },
+    { id: "Qwen/Qwen3.5-9B", name: "Qwen3.5 9B" },
+    { id: "Qwen/Qwen3-8B", name: "Qwen3 8B" },
+    { id: "Qwen/Qwen3-4B-Base", name: "Qwen3 4B Base" },
+    { id: "Qwen/Qwen3-1.7B-Base", name: "Qwen3 1.7B Base" },
+    { id: "meta-llama/Llama-3.1-8B-Instruct", name: "Llama 3.1 8B Instruct" },
+    { id: "meta-llama/Llama-3.2-1B-Instruct", name: "Llama 3.2 1B Instruct" },
+    { id: "google/gemma-3-4b-pt", name: "Gemma 3 4B (Pretrained)" },
+    { id: "HuggingFaceTB/SmolLM3-3B-Base", name: "SmolLM3 3B Base" },
+  ],
+};

src/shared/constants/providers/apikey/frontier-labs.ts

@@ -27,6 +27,23 @@ export const APIKEY_PROVIDERS_FRONTIER = {
     hasFree: true,
     freeNote: "$10/month recurring free API credits",
   },
+  pioneer: {
+    id: "pioneer",
+    alias: "pn",
+    name: "Pioneer AI",
+    icon: "rocket_launch",
+    color: "#7C5CFF",
+    textIcon: "PN",
+    website: "https://pioneer.ai",
+    notice: {
+      text: "Pioneer AI by Fastino Labs. Free $75 usage credits, no credit card required. Use API key auth with a pio_sk_... key. Only open-tier models (Qwen3, Llama, Gemma, SmolLM) work directly — gated models (Claude/GPT/Gemini) require prior fine-tuning via the Pioneer platform.",
+      apiKeyUrl: "https://agent.pioneer.ai/settings/api-keys",
+      signupUrl: "https://agent.pioneer.ai/auth",
+    },
+    hasFree: true,
+    freeNote: "$75 free usage credits — no credit card required",
+    serviceKinds: ["llm"],
+  },
   anthropic: {
     id: "anthropic",
     alias: "anthropic",

tests/unit/pioneer-provider.test.ts

@@ -0,0 +1,83 @@
+import test from "node:test";
+import assert from "node:assert/strict";
+
+import { APIKEY_PROVIDERS, AI_PROVIDERS } from "../../src/shared/constants/providers.ts";
+import { REGISTRY } from "../../open-sse/config/providerRegistry.ts";
+import { PROVIDERS as LEGACY_PROVIDERS } from "../../open-sse/config/constants.ts";
+
+test("pioneer is registered as an API-key provider in the UI catalog", () => {
+  const pioneer = APIKEY_PROVIDERS.pioneer;
+  assert.ok(pioneer, "APIKEY_PROVIDERS.pioneer must exist");
+  assert.equal(pioneer.id, "pioneer");
+  assert.equal(pioneer.alias, "pn");
+  assert.equal(pioneer.name, "Pioneer AI");
+  assert.equal(pioneer.color, "#7C5CFF");
+});
+
+test("pioneer appears in AI_PROVIDERS (composed view)", () => {
+  const pioneer = AI_PROVIDERS.pioneer;
+  assert.ok(pioneer, "AI_PROVIDERS.pioneer must exist");
+  assert.equal(pioneer.id, "pioneer");
+});
+
+test("pioneer registry entry is correct (format, auth, executor)", () => {
+  const pioneer = REGISTRY.pioneer;
+  assert.ok(pioneer, "REGISTRY.pioneer must exist");
+  assert.equal(pioneer.format, "openai");
+  assert.equal(pioneer.executor, "default");
+  assert.equal(pioneer.authType, "apikey");
+  // X-API-Key auth (not Bearer) — matches upstream preference
+  assert.equal(pioneer.authHeader, "x-api-key");
+  assert.equal(pioneer.baseUrl, "https://api.pioneer.ai/v1/chat/completions");
+  assert.equal(pioneer.alias, "pn");
+});
+
+test("pioneer has open-tier models with supports_on_demand_inference", () => {
+  const pioneer = REGISTRY.pioneer;
+  const ids = pioneer.models.map((m) => m.id);
+  // Qwen3 models
+  assert.ok(ids.includes("Qwen/Qwen3-32B"), "must include Qwen3 32B");
+  assert.ok(ids.includes("Qwen/Qwen3-8B"), "must include Qwen3 8B");
+  // Llama models
+  assert.ok(ids.includes("meta-llama/Llama-3.1-8B-Instruct"), "must include Llama 3.1 8B Instruct");
+  assert.ok(ids.includes("meta-llama/Llama-3.2-1B-Instruct"), "must include Llama 3.2 1B Instruct");
+  // At least 10 models total (open-tier catalog)
+  assert.ok(pioneer.models.length >= 10, `expected >= 10 models, got ${pioneer.models.length}`);
+  // Gated models (Claude/GPT/Gemini) must NOT appear — they require fine-tuning first
+  assert.ok(
+    !ids.some((id) => id.toLowerCase().includes("claude")),
+    "gated Claude models must not be in catalog"
+  );
+  assert.ok(
+    !ids.some((id) => id.toLowerCase().includes("gpt")),
+    "gated GPT models must not be in catalog"
+  );
+  assert.ok(
+    !ids.some((id) => id.toLowerCase().includes("gemini")),
+    "gated Gemini models must not be in catalog"
+  );
+});
+
+test("pioneer legacy PROVIDERS entry resolves from generated map", () => {
+  const legacy = LEGACY_PROVIDERS.pioneer;
+  assert.ok(legacy, "LEGACY_PROVIDERS.pioneer must exist (generated from REGISTRY)");
+  assert.equal(legacy.format, "openai");
+  // Should not have OAuth fields — pioneer is API-key only
+  assert.equal(legacy.clientId, undefined);
+  assert.equal(legacy.clientSecret, undefined);
+  assert.equal(legacy.tokenUrl, undefined);
+});
+
+test("pioneer has hasFree flag and free signup notice", () => {
+  const pioneer = APIKEY_PROVIDERS.pioneer;
+  assert.equal(pioneer.hasFree, true, "pioneer should advertise free tier");
+  assert.ok(pioneer.freeNote?.includes("$75"), "freeNote should mention $75 credits");
+  assert.ok(
+    pioneer.notice?.signupUrl?.includes("pioneer.ai"),
+    "signupUrl should point to pioneer.ai"
+  );
+  assert.ok(
+    pioneer.notice?.apiKeyUrl?.includes("pioneer.ai"),
+    "apiKeyUrl should point to pioneer.ai"
+  );
+});