@fudrouter/fsrouter 0.6.64 → 0.6.65

package/open-sse/providers/REGISTRY_TEMPLATE.js

@@ -0,0 +1,98 @@
+/**
+ * REGISTRY ENTRY TEMPLATE — copy into registry/{id}.js when adding a new provider.
+ *
+ * NOT imported by registry/index.js (lives outside registry/, static-import list ignores it).
+ * Delete every block your provider does not need. Only `id` + `category` are required.
+ * Field contract: see schema.js `@typedef RegistryEntry`. Runtime builders: providers/index.js.
+ *
+ * Quick recipes:
+ *   - Plain API-key LLM      → id, alias, category:"apikey", display, transport{baseUrl}, models.
+ *   - OAuth LLM (device/PKCE)→ add oauth{...}; clientId/tokenUrl auto-inject into transport.
+ *   - Media-only (tts/stt/…) → drop `models`+chat baseUrl, fill media{serviceKinds, *Config}.
+ */
+
+// import { CLAUDE_API_HEADERS, GOOGLE_OAUTH_CLIENT, OPENAI_COMPAT_BASE } from "./shared.js";
+
+export default {
+  // ── identity ────────────────────────────────────────────────────────────
+  id: "example",                 // REQUIRED. kebab-case, unique.
+  alias: "ex",                   // short key for PROVIDER_MODELS (defaults to id if omitted).
+  aliases: ["example-ai"],       // optional extra lookup tokens.
+  uiAlias: "ex",                 // optional UI badge token.
+  category: "apikey",            // REQUIRED. "apikey" | "oauth" | "freeTier" | ...
+
+  // ── auth hints (only when relevant) ──────────────────────────────────────
+  authType: "apikey",            // "apikey" | "oauth".
+  hasOAuth: false,               // true if an OAuth flow exists.
+  authModes: ["apikey"],         // e.g. ["oauth","apikey"] when both supported.
+  // noAuth: true,               // local/free providers needing no credential.
+
+  // ── UI display ───────────────────────────────────────────────────────────
+  display: {
+    name: "Example",
+    icon: "bolt",                // material icon name OR textIcon fallback.
+    color: "#3B82F6",
+    textIcon: "EX",
+    website: "https://example.com",
+    notice: { apiKeyUrl: "https://example.com/keys" }, // or signupUrl.
+    // deprecated: true, deprecationNotice: "RISK_NOTICE",
+    // kindNotice: { image: "Requires paid plan." },
+    // mediaPriority: 1,
+  },
+
+  // ── transport (HTTP runtime) → PROVIDERS[id] ─────────────────────────────
+  // Defaults applied: format:"openai". Declare ONLY what differs.
+  transport: {
+    baseUrl: "https://api.example.com/v1/chat/completions",
+    format: "openai",            // "openai" | "claude" | "gemini" | "openai-responses" | ...
+    // validateUrl: "https://api.example.com/v1/models",
+    // headers: { "User-Agent": "..." },           // static fingerprint (anti-ban) lives here.
+    // auth: { header: "x-api-key", scheme: "raw" },
+    // forceStream: true, urlSuffix: "?beta=true",
+    // quirks: { dropOutputConfig: true },
+    // retry: { 429: { attempts: 6 }, 503: { attempts: 3 } },
+    // usage: { url: "https://api.example.com/usage" }, // or { urls: [...] } for multi-call.
+    // modelsFetcher: { url: "https://api.example.com/models", type: "openai" }, // dynamic model list.
+    // regions: { sgp: "https://sgp...", cn: "https://cn..." }, defaultRegion: "sgp",
+    // NOTE: clientId/clientSecret/tokenUrl are injected from `oauth` — do NOT duplicate here.
+  },
+
+  // ── oauth flow → PROVIDER_OAUTH[id] (omit for pure API-key) ───────────────
+  // oauth: {
+  //   clientId: "app_xxx",
+  //   authorizeUrl: "https://auth.example.com/oauth/authorize", // PKCE/code flow.
+  //   tokenUrl: "https://auth.example.com/oauth/token",
+  //   deviceCodeUrl: "https://auth.example.com/device",          // device-code flow.
+  //   refreshUrl: "https://auth.example.com/oauth/token",
+  //   scope: "openid profile offline_access",                    // or scopes: [...].
+  //   codeChallengeMethod: "S256",
+  //   redirectUri: "http://127.0.0.1:1455/auth/callback", fixedPort: 1455, callbackPath: "/auth/callback",
+  //   extraParams: { foo: "bar" },
+  //   refresh: { encoding: "form", scope: "openid offline_access" }, // "form" | "json".
+  //   refreshLeadMs: 300000,
+  //   userInfoUrl: "https://example.com/userinfo",
+  // },
+
+  // ── media (non-LLM services) → PROVIDER_MEDIA[id] ────────────────────────
+  // media: {
+  //   serviceKinds: ["llm", "tts", "stt", "embedding", "image", "imageToText", "webSearch"],
+  //   ttsConfig:       { baseUrl: "...", authType: "apikey", authHeader: "bearer", format: "openai", defaultModel: "tts-1", models: [{ id: "tts-1", name: "TTS-1" }] },
+  //   sttConfig:       { baseUrl: "...", authType: "apikey", authHeader: "bearer", format: "openai", models: [{ id: "whisper-1", name: "Whisper" }] },
+  //   embeddingConfig: { baseUrl: "...", authType: "apikey", authHeader: "bearer", models: [{ id: "emb-1", name: "Emb", dimensions: 1536 }] },
+  //   imageConfig:     { baseUrl: "https://api.example.com/v1/images/generations" },
+  //   searchViaChat:   { defaultModel: "ex-search", pricingUrl: "https://example.com/pricing" },
+  //   // hiddenKinds: ["image"],
+  // },
+
+  // ── models (omit = no key; [] = explicit empty) ──────────────────────────
+  models: [
+    { id: "example-large", name: "Example Large" },
+    // { id: "example-img", name: "Example Image", type: "image", capabilities: ["text2img"], params: ["size"] },
+    // { id: "example-emb", name: "Example Embed", type: "embedding" },
+  ],
+
+  // ── optional flags ───────────────────────────────────────────────────────
+  // features: { usage: true },
+  // thinkingConfig: { options: ["auto", "none", "low", "high"], defaultMode: "auto" },
+  // passthroughModels: true,
+};

package/open-sse/providers/capabilities.js

@@ -0,0 +1,293 @@
+// Model capabilities — what each model can read/do beyond plain text.
+//
+// Fallback order (first match wins), result merged over DEFAULT_CAPABILITIES:
+//   1. PROVIDER_CAPABILITIES[provider][model]  — provider-specific override
+//   2. MODEL_CAPABILITIES[model]               — canonical exact id (handles exceptions)
+//   3. PATTERN_CAPABILITIES                     — glob match, ordered specific -> generic
+//   4. DEFAULT_CAPABILITIES                     — safe floor (always returned)
+//
+// ── HOW TO ADD / UPDATE A MODEL ──────────────────────────────────────
+// Authoritative data source: https://models.dev/api.json (145 providers, 4000+
+// models, MIT). Each model exposes the exact fields we map below:
+//   modalities.input  ["text","image","pdf","audio","video"] -> vision / pdf / audioInput / videoInput
+//   modalities.output ["text","image","audio"]               -> imageOutput / audioOutput
+//   reasoning   -> reasoning      tool_call    -> tools
+//   limit.context -> contextWindow   limit.output -> maxOutput
+// Look up the model id, then:
+//   • If a PATTERN below already covers it correctly -> nothing to do.
+//   • If it is an exception (pattern would mis-match) -> add an exact entry to
+//     MODEL_CAPABILITIES (only the fields that differ from DEFAULT).
+//   • If a whole new family -> add an ordered PATTERN (specific before generic).
+// NOTE: models.dev has NO "search" flag (web search is a runtime tool, not a
+// model spec); set `search` from vendor docs (Claude 4.x+, GPT-5.x/4o, Gemini
+// 2.0+, Grok, Perplexity). Verify with: curl -s https://models.dev/api.json
+
+import { matchPattern } from "./pricing.js";
+
+/**
+ * Safe floor — every resolved result is merged over this so consumers
+ * never need null-checks. Most modern LLMs meet these limits.
+ */
+export const DEFAULT_CAPABILITIES = {
+  // input modalities
+  vision: false,        // read images
+  pdf: false,           // read PDF / documents
+  audioInput: false,    // read audio
+  videoInput: false,    // read video
+  // output modalities
+  imageOutput: false,   // generate images
+  audioOutput: false,   // generate audio
+  // features
+  search: false,        // built-in web search tool / grounding
+  tools: true,          // function / tool calling
+  reasoning: false,     // thinking / reasoning
+  // thinking wire format (only meaningful when reasoning:true). null → derive from transport.format.
+  // enum: openai|claude-adaptive|claude-budget|gemini-level|gemini-budget|zai|qwen|deepseek|kimi|minimax|hunyuan|step
+  thinkingFormat: null,
+  thinkingCanDisable: true,  // false → model cannot turn thinking off (clamp to min instead of disable)
+  thinkingRange: null,       // { min, max } for budget formats; null = no clamp
+  // limits (tokens)
+  contextWindow: 200000,
+  maxOutput: 64000,
+};
+
+// User-added model metadata can carry dashboard service kinds instead of the
+// runtime capability names used here. Map those typed model kinds into input /
+// output capabilities so custom vision models are not treated as text-only.
+const SERVICE_KIND_CAPABILITIES = {
+  imageToText: { vision: true },
+  image: { imageOutput: true },
+  stt: { audioInput: true },
+  tts: { audioOutput: true },
+  embedding: { tools: false },
+};
+
+export function capabilitiesFromServiceKind(kind) {
+  return SERVICE_KIND_CAPABILITIES[kind] || null;
+}
+
+/**
+ * Canonical exact-id overrides — used for exceptions that patterns would
+ * otherwise mis-match. Only declare deltas vs DEFAULT.
+ */
+export const MODEL_CAPABILITIES = {
+  // Claude 4.6/4.7/4.8 and Kiro Sonnet 5 have 1M context + adaptive thinking (override generic claude pattern)
+  "claude-opus-4.6":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4.7":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4-7":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4.8":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4-6":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4-8":   { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4.8-thinking": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-opus-4-8-thinking": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-4.6": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-4-6": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-5": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-5-thinking": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-5-agentic": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+  "claude-sonnet-5-thinking-agentic": { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive", contextWindow: 1000000, maxOutput: 128000 },
+
+  // Gemini image-gen / OpenAI image / xai image variants
+  "gpt-image-1":       { imageOutput: true, tools: false },
+
+  // GLM vision variant (text GLM has no vision)
+  "glm-4.6v":          { vision: true, reasoning: true, thinkingFormat: "zai", contextWindow: 128000 },
+
+  // Qwen plain coder/text (no vision) — registry "vision-model" / "coder-model" aliases
+  "vision-model":      { vision: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000 },
+  "coder-model":       { reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000 },
+};
+
+/**
+ * Provider-specific capability overrides. Keyed by provider alias/id.
+ */
+export const PROVIDER_CAPABILITIES = {
+  // NVIDIA NIM is OpenAI-compatible → rejects MiniMax/GLM native `thinking` field.
+  // Force openai reasoning_effort format for its reasoning models. #issue
+  "nvidia": {
+    "minimaxai/minimax-m2.7": { reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 131072 },
+    "minimaxai/minimax-m3": { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 512000, maxOutput: 131072 },
+    "z-ai/glm-5.2": { reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 128000 },
+    "deepseek-ai/deepseek-v4-pro": { reasoning: true, thinkingFormat: "openai", contextWindow: 1000000, maxOutput: 65536 },
+    "deepseek-ai/deepseek-v4-flash": { reasoning: true, thinkingFormat: "openai", contextWindow: 1000000, maxOutput: 65536 },
+  },
+  // CodeBuddy.cn — authoritative per-model metadata from the gateway's model
+  // config (contextWindow=maxInputTokens, maxOutput=maxOutputTokens, vision=
+  // supportsImages). Every model reasons via OpenAI-style reasoning_effort
+  // (see registry thinkingFormat). `onlyReasoning` models can't turn thinking
+  // off → thinkingCanDisable:false (clamped to minimal instead of disabled).
+  "codebuddy-cn": {
+    "glm-5.2":            { reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 1000000, maxOutput: 48000 },
+    "glm-5.1":            { reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 48000 },
+    "glm-5.0":            { reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 48000 },
+    "glm-5.0-turbo":      { reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 48000 },
+    "glm-5v-turbo":       { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 38000 },
+    "glm-4.7":            { reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 48000 },
+    "minimax-m3":         { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 512000, maxOutput: 48000 },
+    "minimax-m2.7":       { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 48000 },
+    "kimi-k2.7":          { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 256000, maxOutput: 32000 },
+    "kimi-k2.6":          { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 256000, maxOutput: 32000 },
+    "kimi-k2.5":          { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 164000, maxOutput: 32000 },
+    "hy3-preview":        { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 192000, maxOutput: 64000 },
+    "deepseek-v4-pro":    { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 1000000, maxOutput: 50000 },
+    "deepseek-v4-flash":  { vision: true, reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 1000000, maxOutput: 50000 },
+    "deepseek-v3-2-volc": { reasoning: true, thinkingFormat: "openai", thinkingCanDisable: false, contextWindow: 96000, maxOutput: 32000 },
+  },
+};
+
+/**
+ * Pattern fallback — glob (* = wildcard), matched case-insensitively and
+ * anchored (^...$) so a pattern must match the full model id. ORDER MATTERS:
+ * vision/specific variants first, text-only/generic families last, to avoid
+ * a broad family pattern swallowing an exception (e.g. glm-4.6v vs glm-5).
+ */
+export const PATTERN_CAPABILITIES = [
+  // ── Claude (4.6+ = adaptive thinking; older/haiku = budget) ──────
+  { pattern: "*claude*opus-4.6*",   caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive" } },
+  { pattern: "*claude*opus-4.7*",   caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive" } },
+  { pattern: "*claude*opus-4.8*",   caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive" } },
+  { pattern: "*claude*sonnet-4.6*", caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive" } },
+  { pattern: "*claude*sonnet-4.7*", caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-adaptive" } },
+  { pattern: "*claude*haiku*",  caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget" } },
+  { pattern: "*claude*opus*",   caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget" } },
+  { pattern: "*claude*sonnet*", caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget" } },
+  { pattern: "*claude*fable*",  caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget", contextWindow: 1000000, maxOutput: 128000 } },
+  { pattern: "*claude*mythos*", caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget", contextWindow: 1000000, maxOutput: 128000 } },
+  { pattern: "*claude-3*",      caps: { vision: true } },
+  { pattern: "*claude*",        caps: { vision: true, reasoning: true, search: true, thinkingFormat: "claude-budget" } },
+
+  // ── Gemini (all 2.0+ multimodal + google_search grounding, 1M ctx) ─
+  { pattern: "*gemini*image*",  caps: { vision: true, imageOutput: true, contextWindow: 1048576 } },
+  { pattern: "*gemini-3*pro*",  caps: { vision: true, audioInput: true, videoInput: true, reasoning: true, search: true, thinkingFormat: "gemini-level", thinkingCanDisable: false, contextWindow: 1048576, maxOutput: 65535 } },
+  { pattern: "*gemini-3*",      caps: { vision: true, audioInput: true, videoInput: true, reasoning: true, search: true, thinkingFormat: "gemini-level", thinkingCanDisable: false, contextWindow: 1048576, maxOutput: 65536 } },
+  { pattern: "*gemini-2.5*",    caps: { vision: true, audioInput: true, videoInput: true, reasoning: true, search: true, thinkingFormat: "gemini-budget", thinkingRange: { min: 0, max: 24576 }, contextWindow: 1048576, maxOutput: 65536 } },
+  { pattern: "*gemini-2*",      caps: { vision: true, audioInput: true, videoInput: true, search: true, contextWindow: 1048576, maxOutput: 65536 } },
+  { pattern: "*gemini*",        caps: { vision: true, search: true, contextWindow: 1048576 } },
+  { pattern: "*gemma*",         caps: { vision: true, contextWindow: 128000 } },
+  { pattern: "*nanobanana*",    caps: { vision: true, imageOutput: true } },
+
+  // ── OpenAI GPT-5.x (vision + thinking + web search) ──────────────
+  { pattern: "*gpt-5*image*",   caps: { imageOutput: true } },
+  { pattern: "*gpt-5*codex*",   caps: { reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 400000, maxOutput: 128000 } },
+  { pattern: "*gpt-5*",         caps: { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 400000, maxOutput: 128000 } },
+  { pattern: "*gpt-4o*",        caps: { vision: true, search: true, contextWindow: 128000, maxOutput: 16384 } },
+  { pattern: "*gpt-4.1*",       caps: { vision: true, contextWindow: 1000000, maxOutput: 32768 } },
+  { pattern: "*gpt-4-turbo*",   caps: { vision: true, contextWindow: 128000 } },
+  { pattern: "*gpt-4*",         caps: { contextWindow: 128000 } },
+  { pattern: "*gpt-3.5*",       caps: { contextWindow: 16385, maxOutput: 4096 } },
+  { pattern: "*gpt-oss*",       caps: { reasoning: true, thinkingFormat: "openai", contextWindow: 128000 } },
+
+  // ── OpenAI o-series (reasoning, vision) ──────────────────────────
+  { pattern: "*o1-mini*",       caps: { reasoning: true, thinkingFormat: "openai", contextWindow: 128000 } },
+  { pattern: "*o1*",            caps: { vision: true, reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 100000 } },
+  { pattern: "*o3*",            caps: { vision: true, reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 100000 } },
+  { pattern: "*o4*",            caps: { vision: true, reasoning: true, thinkingFormat: "openai", contextWindow: 200000, maxOutput: 100000 } },
+
+  // ── Grok (vision + Live Search) ──────────────────────────────────
+  { pattern: "*grok*image*",    caps: { imageOutput: true } },
+  { pattern: "*grok-code*",     caps: { reasoning: true, thinkingFormat: "openai", contextWindow: 256000 } },
+  // Grok 4.5 (Grok CLI / Grok Build): 500k context per cli-chat-proxy /v1/models
+  { pattern: "*grok-4.5*",      caps: { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 500000, maxOutput: 64000 } },
+  { pattern: "*grok-4*",        caps: { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 256000 } },
+  { pattern: "*grok-3*",        caps: { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 131072 } },
+  { pattern: "*grok*",          caps: { vision: true, reasoning: true, search: true, thinkingFormat: "openai", contextWindow: 256000 } },
+
+  // ── Qwen (3.5+ = native vision/video; coder & max = text-only; QwQ = thinking-only) ─
+  { pattern: "*qwen*vl*",       caps: { vision: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 262144 } },
+  { pattern: "*qwen*omni*",     caps: { vision: true, audioInput: true, videoInput: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 262144, maxOutput: 65536 } },
+  { pattern: "*qwen*coder*",    caps: { reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000 } },
+  { pattern: "*qwen*max*",      caps: { reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000, maxOutput: 65536 } },
+  { pattern: "*qwen3.5*",       caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000, maxOutput: 65536 } },
+  { pattern: "*qwen3.6*",       caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000, maxOutput: 65536 } },
+  { pattern: "*qwen3.7*",       caps: { vision: true, videoInput: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000, maxOutput: 65536 } },
+  { pattern: "*qwen*plus*",     caps: { vision: true, reasoning: true, thinkingFormat: "qwen", contextWindow: 1000000, maxOutput: 65536 } },
+  { pattern: "*qwen*235b*",     caps: { reasoning: true, thinkingFormat: "qwen", contextWindow: 262144 } },
+  { pattern: "*qwq*",           caps: { reasoning: true, thinkingFormat: "qwen", thinkingCanDisable: false, contextWindow: 131072 } },
+  { pattern: "*qwen*",          caps: { reasoning: true, thinkingFormat: "qwen", contextWindow: 262144 } },
+
+  // ── Kimi (enabled→reasoning_effort; K2.7-code cannot disable) ─────
+  { pattern: "*kimi*k2.7*code*", caps: { vision: true, reasoning: true, thinkingFormat: "kimi", thinkingCanDisable: false, contextWindow: 262144, maxOutput: 262144 } },
+  { pattern: "*kimi*k2*",       caps: { vision: true, reasoning: true, thinkingFormat: "kimi", contextWindow: 262144, maxOutput: 262144 } },
+  { pattern: "*kimi*",          caps: { reasoning: true, thinkingFormat: "kimi", contextWindow: 262144 } },
+
+  // ── GLM / Z.ai (thinking.enabled; disable via enable_thinking:false) ─
+  { pattern: "*glm-5*",         caps: { reasoning: true, thinkingFormat: "zai", contextWindow: 200000, maxOutput: 128000 } },
+  { pattern: "*glm-4.7*",       caps: { reasoning: true, thinkingFormat: "zai", contextWindow: 200000, maxOutput: 128000 } },
+  { pattern: "*glm-4*",         caps: { reasoning: true, thinkingFormat: "zai", contextWindow: 200000 } },
+  { pattern: "*glm*",           caps: { reasoning: true, thinkingFormat: "zai", contextWindow: 200000 } },
+
+  // ── DeepSeek (thinking.enabled + reasoning_effort; r1 = thinking-only) ─
+  { pattern: "*deepseek-v4*",   caps: { reasoning: true, thinkingFormat: "deepseek", contextWindow: 1000000, maxOutput: 384000 } },
+  { pattern: "*reasoner*",      caps: { reasoning: true, thinkingFormat: "deepseek", thinkingCanDisable: false, contextWindow: 128000 } },
+  { pattern: "*deepseek-r*",    caps: { reasoning: true, thinkingFormat: "deepseek", thinkingCanDisable: false, contextWindow: 128000 } },
+  { pattern: "*deepseek-chat*", caps: { contextWindow: 128000 } },
+  { pattern: "*deepseek*",      caps: { reasoning: true, thinkingFormat: "deepseek", contextWindow: 128000 } },
+
+  // ── MiniMax (M3 = adaptive; M2.x cannot disable) ─────────────────
+  { pattern: "*minimax*image*", caps: { imageOutput: true } },
+  { pattern: "*minimax-m3*",    caps: { vision: true, reasoning: true, thinkingFormat: "minimax", contextWindow: 1048576, maxOutput: 512000 } },
+  { pattern: "*minimax-m2.7*",  caps: { reasoning: true, thinkingFormat: "minimax", thinkingCanDisable: false, contextWindow: 204800, maxOutput: 131072 } },
+  { pattern: "*minimax*",       caps: { reasoning: true, thinkingFormat: "minimax", thinkingCanDisable: false, contextWindow: 200000, maxOutput: 131072 } },
+
+  // ── Xiaomi MiMo (vision, 1M / 262K ctx) ──────────────────────────
+  { pattern: "*mimo*v2.5*",     caps: { vision: true, contextWindow: 1048576, maxOutput: 131072 } },
+  { pattern: "*mimo*omni*",     caps: { vision: true, audioInput: true, contextWindow: 262144, maxOutput: 131072 } },
+  { pattern: "*mimo*",          caps: { vision: true, contextWindow: 262144, maxOutput: 131072 } },
+
+  // ── Llama (4 = vision/1M; 3.x = text-only/128K) ──────────────────
+  { pattern: "*llama-4*",       caps: { vision: true, contextWindow: 1000000 } },
+  { pattern: "*llama*",         caps: { contextWindow: 128000 } },
+
+  // ── Mistral (Large 3 = vision/256K; codestral text) ──────────────
+  { pattern: "*codestral*",     caps: { contextWindow: 256000 } },
+  { pattern: "*mistral-large*", caps: { vision: true, contextWindow: 256000 } },
+  { pattern: "*mistral*",       caps: { contextWindow: 128000 } },
+
+  // ── Cohere (Command A Vision = vision; others text) ──────────────
+  { pattern: "*command-a-vision*", caps: { vision: true, contextWindow: 128000 } },
+  { pattern: "*command*",       caps: { contextWindow: 128000 } },
+
+  // ── Perplexity (web search native) ───────────────────────────────
+  { pattern: "*sonar*",         caps: { search: true, contextWindow: 128000 } },
+  { pattern: "*pplx*",          caps: { search: true, contextWindow: 128000 } },
+  { pattern: "*perplexity*",    caps: { search: true, contextWindow: 128000 } },
+
+  // ── Others ───────────────────────────────────────────────────────
+  { pattern: "*hunyuan*",       caps: { reasoning: true, thinkingFormat: "hunyuan", contextWindow: 262144, maxOutput: 262144 } },
+  { pattern: "hy3*",            caps: { reasoning: true, thinkingFormat: "hunyuan", contextWindow: 262144, maxOutput: 262144 } },
+  { pattern: "*step-*",         caps: { reasoning: true, thinkingFormat: "step", contextWindow: 128000 } },
+  { pattern: "*nemotron*",      caps: { reasoning: true, contextWindow: 128000 } },
+  { pattern: "*ling-*",         caps: { reasoning: true, contextWindow: 128000 } },
+];
+
+/**
+ * Resolve capabilities for a model using the 4-step fallback chain,
+ * merged over DEFAULT_CAPABILITIES so the result is always complete.
+ *
+ * @param {string} provider
+ * @param {string} model
+ * @returns {object} full capabilities object
+ */
+export function getCapabilitiesForModel(provider, model) {
+  if (!model) return { ...DEFAULT_CAPABILITIES };
+
+  // 1. Provider-specific override
+  if (provider && PROVIDER_CAPABILITIES[provider]?.[model]) {
+    return { ...DEFAULT_CAPABILITIES, ...PROVIDER_CAPABILITIES[provider][model] };
+  }
+
+  // 2. Canonical exact (strip vendor prefix: "anthropic/claude-opus-4.7" -> "claude-opus-4.7")
+  const baseModel = model.includes("/") ? model.split("/").pop() : model;
+  if (MODEL_CAPABILITIES[baseModel]) return { ...DEFAULT_CAPABILITIES, ...MODEL_CAPABILITIES[baseModel] };
+  if (MODEL_CAPABILITIES[model]) return { ...DEFAULT_CAPABILITIES, ...MODEL_CAPABILITIES[model] };
+
+  // 3. Pattern match (first match wins)
+  for (const { pattern, caps } of PATTERN_CAPABILITIES) {
+    if (matchPattern(pattern, baseModel) || matchPattern(pattern, model)) {
+      return { ...DEFAULT_CAPABILITIES, ...caps };
+    }
+  }
+
+  // 4. Floor
+  return { ...DEFAULT_CAPABILITIES };
+}

package/open-sse/providers/index.js

@@ -0,0 +1,51 @@
+// Single source: build PROVIDERS + PROVIDER_MODELS from registry/{id}.js (transport + models co-located).
+import REGISTRY from "./registry/index.js";
+import { PROVIDER_DEFAULTS } from "./schema.js";
+import { normalizeModel } from "./models/schema.js";
+import { buildTtsProviderModels } from "../config/ttsModels.js";
+
+// oauth block is canonical for these fields; inject into transport so executors reading
+// this.config.{clientId,clientSecret,tokenUrl} keep working without duplicating in transport
+const OAUTH_INJECT_FIELDS = ["clientId", "clientSecret", "tokenUrl"];
+
+// transport: re-apply shared default (format:"openai") + inject oauth-canonical fields
+function buildTransport(transport, oauth) {
+  const t = { ...transport };
+  if (!t.format) t.format = PROVIDER_DEFAULTS.format;
+  if (oauth) {
+    for (const f of OAUTH_INJECT_FIELDS) {
+      if (t[f] === undefined && oauth[f] !== undefined) t[f] = oauth[f];
+    }
+  }
+  return t;
+}
+
+const MEDIA_KEYS = new Set([
+  "serviceKinds", "ttsConfig", "sttConfig", "embeddingConfig",
+  "imageConfig", "imageToTextConfig", "videoConfig", "musicConfig",
+  "searchViaChat", "searchConfig", "fetchConfig",
+  "modelsFetcher", "mediaPriority", "hiddenKinds",
+]);
+
+export const PROVIDERS = {};
+export const PROVIDER_MODELS = {};
+export const PROVIDER_OAUTH = {};
+export const PROVIDER_MEDIA = {};
+for (const entry of REGISTRY) {
+  if (entry.transport) {
+    PROVIDERS[entry.id] = buildTransport(entry.transport, entry.oauth);
+    if (entry.transports) PROVIDERS[entry.id].transports = entry.transports;
+  }
+  if (entry.models !== undefined) PROVIDER_MODELS[entry.alias || entry.id] = entry.models.map(normalizeModel);
+  if (entry.oauth) PROVIDER_OAUTH[entry.id] = entry.oauth;
+  // Build PROVIDER_MEDIA from top-level fields (post-migration) + legacy entry.media
+  const mediaFields = {};
+  for (const k of MEDIA_KEYS) {
+    if (entry[k] !== undefined) mediaFields[k] = entry[k];
+  }
+  if (entry.media) Object.assign(mediaFields, entry.media);
+  if (Object.keys(mediaFields).length) PROVIDER_MEDIA[entry.id] = mediaFields;
+}
+
+// TTS model/voice tables keyed by special names (openai-tts-models, ...), not provider ids
+Object.assign(PROVIDER_MODELS, buildTtsProviderModels());

package/open-sse/providers/models/helpers.js

@@ -0,0 +1,20 @@
+// Codex auto-generates a "-review" variant for each llm model (review quota family)
+export const CODEX_REVIEW_SUFFIX = "-review";
+
+export function withCodexReviewModels(models) {
+  return models.flatMap((model) => {
+    if ((model.kind || model.type || "llm") !== "llm" || model.id.endsWith(CODEX_REVIEW_SUFFIX)) {
+      return [model];
+    }
+    return [
+      model,
+      {
+        ...model,
+        id: `${model.id}${CODEX_REVIEW_SUFFIX}`,
+        name: `${model.name} Review`,
+        upstreamModelId: model.upstreamModelId || model.id,
+        quotaFamily: "review"
+      }
+    ];
+  });
+}

package/open-sse/providers/models/namePatterns.js

@@ -0,0 +1,33 @@
+// Derive a display name from a model id when the entry omits `name` (mirrors PATTERN_PRICING).
+// Provider entries that ship their own `name` always win; this is only a fallback for terse entries.
+
+// Capitalize a hyphen/space separated token group: "coder-plus" → "Coder Plus".
+function titleCase(s) {
+  return s
+    .split(/[-_\s]+/)
+    .filter(Boolean)
+    .map((w) => (/^\d/.test(w) ? w : w.charAt(0).toUpperCase() + w.slice(1)))
+    .join(" ");
+}
+
+// Ordered: first match wins. Keep specific patterns above generic ones.
+export const NAME_PATTERNS = [
+  [/^kimi-k(\d+(?:\.\d+)?)(-thinking)?$/i, (m) => `Kimi K${m[1]}${m[2] ? " Thinking" : ""}`],
+  [/^glm-(\d+(?:\.\d+)?)(v)?$/i, (m) => `GLM ${m[1]}${m[2] ? "V (Vision)" : ""}`],
+  [/^minimax-m(\d+(?:\.\d+)?)$/i, (m) => `MiniMax M${m[1]}`],
+  [/^gpt-(.+)$/i, (m) => `GPT ${titleCase(m[1])}`],
+  [/^gemini-(.+)$/i, (m) => `Gemini ${titleCase(m[1])}`],
+  [/^grok-(.+)$/i, (m) => `Grok ${titleCase(m[1])}`],
+  [/^deepseek-(.+)$/i, (m) => `DeepSeek ${titleCase(m[1])}`],
+  [/^qwen([\d.]+.*)$/i, (m) => `Qwen ${titleCase(m[1])}`],
+];
+
+// id → display name (regex fallback → id verbatim)
+export function deriveModelName(id) {
+  if (typeof id !== "string") return id;
+  for (const [re, fn] of NAME_PATTERNS) {
+    const m = id.match(re);
+    if (m) return fn(m);
+  }
+  return id;
+}

package/open-sse/providers/models/schema.js

@@ -0,0 +1,40 @@
+import { deriveModelName } from "./namePatterns.js";
+
+// Normalize version separators in a model id: hyphen between two digits becomes a dot.
+// Registry ids use dots for versions ("claude-sonnet-4.5") but clients (CLIs, aliases)
+// often send them with dashes ("claude-sonnet-4-5"). Only digit-digit hyphens are
+// touched, so word/suffix hyphens stay intact ("-thinking", "-agentic", "qwen3-coder-next").
+export function normalizeModelId(modelId) {
+  if (typeof modelId !== "string") return modelId;
+  return modelId.replace(/(\d)-(\d)/g, "$1.$2");
+}
+
+// Model defaults centralized (was scattered as `m.kind || "llm"`, `quotaFamily || "normal"`, etc.)
+export const MODEL_DEFAULTS = {
+  kind: "llm",
+  quotaFamily: "normal",
+  strip: [],
+  targetFormat: null
+};
+
+// Normalize a registry model entry: accept terse "id" string, fill name via regex when omitted.
+// Override always wins (raw spread last); name falls back to regex → id.
+export function normalizeModel(raw) {
+  const model = typeof raw === "string" ? { id: raw } : raw;
+  if (model.name !== undefined) return model;
+  return { ...model, name: deriveModelName(model.id) };
+}
+
+// Resolve model kind with default (accepts legacy `type` field)
+export function modelKind(model) {
+  return model?.kind || model?.type || MODEL_DEFAULTS.kind;
+}
+export function modelQuotaFamily(model) {
+  return model?.quotaFamily || MODEL_DEFAULTS.quotaFamily;
+}
+export function modelStrip(model) {
+  return model?.strip || [];
+}
+export function modelTargetFormat(model) {
+  return model?.targetFormat || MODEL_DEFAULTS.targetFormat;
+}