claudish 5.7.1 → 5.9.0

package/dist/index.js

@@ -29458,7 +29458,7 @@ var getRemoteProviders = () => [
     authScheme: "bearer",
     capabilities: {
       supportsTools: true,
-      supportsVision: true,
+      supportsVision: false,
       supportsStreaming: true,
       supportsJsonMode: false,
       supportsReasoning: false
@@ -29941,9 +29941,7 @@ function checkApiKeyForProvider(nativeProvider, modelName) {
 }
 function getAutoRouteHint(modelName, nativeProvider) {
   const hint = PROVIDER_HINT_MAP[nativeProvider];
-  const lines = [
-    `No credentials found for "${modelName}". Options:`
-  ];
+  const lines = [`No credentials found for "${modelName}". Options:`];
   let hasOption = false;
   if (hint?.loginFlag) {
     lines.push(`  Run:  claudish ${hint.loginFlag}  (authenticate via OAuth)`);
@@ -30004,7 +30002,61 @@ function autoRoute(modelName, nativeProvider) {
   }
   return null;
 }
-var API_KEY_ENV_VARS, PROVIDER_HINT_MAP;
+function hasProviderCredentials(provider) {
+  const keyInfo = API_KEY_ENV_VARS[provider];
+  if (keyInfo?.envVar && process.env[keyInfo.envVar])
+    return true;
+  if (keyInfo?.aliases?.some((a) => process.env[a]))
+    return true;
+  return hasOAuthCredentials(provider);
+}
+function getFallbackChain(modelName, nativeProvider) {
+  const routes = [];
+  const litellmBaseUrl = process.env.LITELLM_BASE_URL;
+  if (litellmBaseUrl && process.env.LITELLM_API_KEY) {
+    routes.push({
+      provider: "litellm",
+      modelSpec: `litellm@${modelName}`,
+      displayName: "LiteLLM"
+    });
+  }
+  if (process.env.OPENCODE_API_KEY) {
+    routes.push({
+      provider: "opencode-zen",
+      modelSpec: `zen@${modelName}`,
+      displayName: "OpenCode Zen"
+    });
+  }
+  const sub = SUBSCRIPTION_ALTERNATIVES[nativeProvider];
+  if (sub && hasProviderCredentials(sub.subscriptionProvider)) {
+    const subModelName = sub.modelName || modelName;
+    routes.push({
+      provider: sub.subscriptionProvider,
+      modelSpec: `${sub.prefix}@${subModelName}`,
+      displayName: sub.displayName
+    });
+  }
+  if (nativeProvider !== "unknown" && nativeProvider !== "qwen" && nativeProvider !== "native-anthropic") {
+    if (hasProviderCredentials(nativeProvider)) {
+      const prefix = PROVIDER_TO_PREFIX[nativeProvider] || nativeProvider;
+      routes.push({
+        provider: nativeProvider,
+        modelSpec: `${prefix}@${modelName}`,
+        displayName: DISPLAY_NAMES[nativeProvider] || nativeProvider
+      });
+    }
+  }
+  if (process.env.OPENROUTER_API_KEY) {
+    const resolution = resolveModelNameSync(modelName, "openrouter");
+    routes.push({
+      provider: "openrouter",
+      modelSpec: resolution.resolvedId,
+      displayName: "OpenRouter"
+    });
+  }
+  return routes;
+}
+var API_KEY_ENV_VARS, PROVIDER_HINT_MAP, PROVIDER_TO_PREFIX, DISPLAY_NAMES, SUBSCRIPTION_ALTERNATIVES;
 var init_auto_route = __esm(() => {
   init_oauth_registry();
   init_model_catalog_resolver();
@@ -30068,6 +30120,65 @@ var init_auto_route = __esm(() => {
       apiKeyEnvVar: "OLLAMA_API_KEY"
     }
   };
+  PROVIDER_TO_PREFIX = {
+    google: "g",
+    openai: "oai",
+    minimax: "mm",
+    "minimax-coding": "mmc",
+    kimi: "kimi",
+    "kimi-coding": "kc",
+    glm: "glm",
+    "glm-coding": "gc",
+    zai: "zai",
+    ollamacloud: "oc",
+    "opencode-zen": "zen",
+    "opencode-zen-go": "zengo",
+    litellm: "ll",
+    vertex: "v",
+    "gemini-codeassist": "go"
+  };
+  DISPLAY_NAMES = {
+    google: "Gemini",
+    openai: "OpenAI",
+    minimax: "MiniMax",
+    "minimax-coding": "MiniMax Coding",
+    kimi: "Kimi",
+    "kimi-coding": "Kimi Coding",
+    glm: "GLM",
+    "glm-coding": "GLM Coding",
+    zai: "Z.AI",
+    ollamacloud: "OllamaCloud",
+    "opencode-zen": "OpenCode Zen",
+    "opencode-zen-go": "OpenCode Zen Go",
+    litellm: "LiteLLM",
+    openrouter: "OpenRouter"
+  };
+  SUBSCRIPTION_ALTERNATIVES = {
+    kimi: {
+      subscriptionProvider: "kimi-coding",
+      modelName: "kimi-for-coding",
+      prefix: "kc",
+      displayName: "Kimi Coding"
+    },
+    minimax: {
+      subscriptionProvider: "minimax-coding",
+      modelName: null,
+      prefix: "mmc",
+      displayName: "MiniMax Coding"
+    },
+    glm: {
+      subscriptionProvider: "glm-coding",
+      modelName: null,
+      prefix: "gc",
+      displayName: "GLM Coding"
+    },
+    google: {
+      subscriptionProvider: "gemini-codeassist",
+      modelName: null,
+      prefix: "go",
+      displayName: "Gemini Code Assist"
+    }
+  };
 });
 
 // src/providers/provider-resolver.ts
@@ -30498,7 +30609,15 @@ __export(exports_cli, {
   getMissingKeyResolutions: () => getMissingKeyResolutions,
   getMissingKeyError: () => getMissingKeyError
 });
-import { readFileSync as readFileSync10, writeFileSync as writeFileSync5, existsSync as existsSync12, mkdirSync as mkdirSync5, copyFileSync, readdirSync, unlinkSync as unlinkSync3 } from "node:fs";
+import {
+  readFileSync as readFileSync10,
+  writeFileSync as writeFileSync5,
+  existsSync as existsSync12,
+  mkdirSync as mkdirSync5,
+  copyFileSync,
+  readdirSync,
+  unlinkSync as unlinkSync3
+} from "node:fs";
 import { fileURLToPath as fileURLToPath3 } from "node:url";
 import { dirname as dirname3, join as join12 } from "node:path";
 import { homedir as homedir11 } from "node:os";
@@ -31836,7 +31955,7 @@ async function fetchGLMCodingModels() {
     return [];
   }
 }
-var __filename4, __dirname4, VERSION = "5.7.1", CACHE_MAX_AGE_DAYS2 = 2, MODELS_JSON_PATH, CLAUDISH_CACHE_DIR2, ALL_MODELS_JSON_PATH;
+var __filename4, __dirname4, VERSION = "5.9.0", CACHE_MAX_AGE_DAYS2 = 2, MODELS_JSON_PATH, CLAUDISH_CACHE_DIR2, ALL_MODELS_JSON_PATH;
 var init_cli = __esm(() => {
   init_config();
   init_model_loader();
@@ -34501,7 +34620,9 @@ async function fetchZenGoModels() {
     return [];
   const ZEN_GO_BASE = process.env.OPENCODE_BASE_URL ? process.env.OPENCODE_BASE_URL.replace("/zen", "/zen/go") : "https://opencode.ai/zen/go";
   try {
-    const mdevResp = await fetch("https://models.dev/api.json", { signal: AbortSignal.timeout(5000) });
+    const mdevResp = await fetch("https://models.dev/api.json", {
+      signal: AbortSignal.timeout(5000)
+    });
     if (!mdevResp.ok)
       return [];
     const mdevData = await mdevResp.json();
@@ -34513,7 +34634,11 @@ async function fetchZenGoModels() {
         const r = await fetch(`${ZEN_GO_BASE}/v1/chat/completions`, {
           method: "POST",
           headers: { "Content-Type": "application/json", Authorization: `Bearer ${apiKey}` },
-          body: JSON.stringify({ model: modelId, messages: [{ role: "user", content: "hi" }], max_tokens: 1 }),
+          body: JSON.stringify({
+            model: modelId,
+            messages: [{ role: "user", content: "hi" }],
+            max_tokens: 1
+          }),
           signal: AbortSignal.timeout(8000)
         });
         if (!r.ok)
@@ -34974,7 +35099,10 @@ async function getAllModelsForSearch(forceUpdate = false) {
   const litellmBaseUrl = process.env.LITELLM_BASE_URL;
   const litellmApiKey = process.env.LITELLM_API_KEY;
   const fetchEntries = [
-    { name: "OpenRouter", promise: fetchAllModels(forceUpdate).then((models) => models.map(toModelInfo)) },
+    {
+      name: "OpenRouter",
+      promise: fetchAllModels(forceUpdate).then((models) => models.map(toModelInfo))
+    },
     { name: "xAI", promise: fetchXAIModels() },
     { name: "Gemini", promise: fetchGeminiModels() },
     { name: "OpenAI", promise: fetchOpenAIModels() },
@@ -34985,7 +35113,10 @@ async function getAllModelsForSearch(forceUpdate = false) {
     { name: "Zen Go", promise: fetchZenGoModels() }
   ];
   if (litellmBaseUrl && litellmApiKey) {
-    fetchEntries.push({ name: "LiteLLM", promise: fetchLiteLLMModels(litellmBaseUrl, litellmApiKey, forceUpdate) });
+    fetchEntries.push({
+      name: "LiteLLM",
+      promise: fetchLiteLLMModels(litellmBaseUrl, litellmApiKey, forceUpdate)
+    });
   }
   const settled = await Promise.allSettled(fetchEntries.map((e) => e.promise));
   const fetchResults = {};
@@ -35197,9 +35328,24 @@ function getKnownModels(provider) {
       { id: "google@gemini-2.0-flash", name: "Gemini 2.0 Flash", context: "1M" }
     ],
     openai: [
-      { id: "oai@gpt-5.3-codex", name: "GPT-5.3 Codex", context: "400K", description: "Latest coding model" },
-      { id: "oai@gpt-5.2-codex", name: "GPT-5.2 Codex", context: "400K", description: "Coding model" },
-      { id: "oai@gpt-5.1-codex-mini", name: "GPT-5.1 Codex Mini", context: "400K", description: "Fast coding model" },
+      {
+        id: "oai@gpt-5.3-codex",
+        name: "GPT-5.3 Codex",
+        context: "400K",
+        description: "Latest coding model"
+      },
+      {
+        id: "oai@gpt-5.2-codex",
+        name: "GPT-5.2 Codex",
+        context: "400K",
+        description: "Coding model"
+      },
+      {
+        id: "oai@gpt-5.1-codex-mini",
+        name: "GPT-5.1 Codex Mini",
+        context: "400K",
+        description: "Fast coding model"
+      },
       { id: "oai@o3", name: "o3", context: "200K", description: "Reasoning model" },
       { id: "oai@o4-mini", name: "o4-mini", context: "200K", description: "Fast reasoning model" },
       { id: "oai@gpt-4.1", name: "GPT-4.1", context: "1M", description: "Large context model" }
@@ -35207,39 +35353,102 @@ function getKnownModels(provider) {
     xai: [
       { id: "xai@grok-4", name: "Grok 4", context: "256K" },
       { id: "xai@grok-4-fast", name: "Grok 4 Fast", context: "2M" },
-      { id: "xai@grok-code-fast-1", name: "Grok Code Fast 1", context: "256K", description: "Optimized for coding" }
+      {
+        id: "xai@grok-code-fast-1",
+        name: "Grok Code Fast 1",
+        context: "256K",
+        description: "Optimized for coding"
+      }
     ],
     minimax: [
-      { id: "mm@minimax-m2.1", name: "MiniMax M2.1", context: "196K", description: "Lightweight coding model" }
+      {
+        id: "mm@minimax-m2.1",
+        name: "MiniMax M2.1",
+        context: "196K",
+        description: "Lightweight coding model"
+      }
     ],
     "minimax-coding": [
-      { id: "mmc@minimax-m2.5", name: "MiniMax M2.5", context: "196K", description: "MiniMax Coding subscription model" },
-      { id: "mmc@minimax-m2.1", name: "MiniMax M2.1", context: "196K", description: "MiniMax Coding subscription model" }
+      {
+        id: "mmc@minimax-m2.5",
+        name: "MiniMax M2.5",
+        context: "196K",
+        description: "MiniMax Coding subscription model"
+      },
+      {
+        id: "mmc@minimax-m2.1",
+        name: "MiniMax M2.1",
+        context: "196K",
+        description: "MiniMax Coding subscription model"
+      }
     ],
     kimi: [
       { id: "kimi@kimi-k2-thinking-turbo", name: "Kimi K2 Thinking Turbo", context: "128K" },
       { id: "kimi@moonshot-v1-128k", name: "Moonshot V1 128K", context: "128K" }
     ],
     "kimi-coding": [
-      { id: "kc@kimi-for-coding", name: "Kimi for Coding", context: "128K", description: "Kimi Coding subscription model" }
+      {
+        id: "kc@kimi-for-coding",
+        name: "Kimi for Coding",
+        context: "128K",
+        description: "Kimi Coding subscription model"
+      }
     ],
     glm: [
-      { id: "glm@glm-5", name: "GLM-5", context: "200K", description: "Latest GLM model with reasoning" },
-      { id: "glm@glm-4.7", name: "GLM-4.7", context: "200K", description: "GLM 4.7 with reasoning" },
-      { id: "glm@glm-4.7-flash", name: "GLM-4.7 Flash", context: "200K", description: "Fast GLM 4.7" },
+      {
+        id: "glm@glm-5",
+        name: "GLM-5",
+        context: "200K",
+        description: "Latest GLM model with reasoning"
+      },
+      {
+        id: "glm@glm-4.7",
+        name: "GLM-4.7",
+        context: "200K",
+        description: "GLM 4.7 with reasoning"
+      },
+      {
+        id: "glm@glm-4.7-flash",
+        name: "GLM-4.7 Flash",
+        context: "200K",
+        description: "Fast GLM 4.7"
+      },
       { id: "glm@glm-4.6", name: "GLM-4.6", context: "200K" },
       { id: "glm@glm-4.5-flash", name: "GLM-4.5 Flash", context: "128K" }
     ],
-    zai: [
-      { id: "zai@glm-4.7", name: "GLM 4.7 (Z.AI)", context: "128K" }
-    ],
+    zai: [{ id: "zai@glm-4.7", name: "GLM 4.7 (Z.AI)", context: "128K" }],
     ollamacloud: [
       { id: "oc@glm-5", name: "GLM-5", context: "203K", description: "GLM-5 on OllamaCloud" },
-      { id: "oc@deepseek-v3.2", name: "DeepSeek V3.2", context: "164K", description: "DeepSeek V3.2 on OllamaCloud" },
-      { id: "oc@gemini-3-pro-preview", name: "Gemini 3 Pro Preview", context: "1M", description: "Gemini 3 Pro on OllamaCloud" },
-      { id: "oc@kimi-k2.5", name: "Kimi K2.5", context: "262K", description: "Kimi K2.5 on OllamaCloud" },
-      { id: "oc@qwen3-coder-next", name: "Qwen3 Coder Next", context: "262K", description: "Qwen3 Coder on OllamaCloud" },
-      { id: "oc@minimax-m2.1", name: "MiniMax M2.1", context: "205K", description: "MiniMax M2.1 on OllamaCloud" }
+      {
+        id: "oc@deepseek-v3.2",
+        name: "DeepSeek V3.2",
+        context: "164K",
+        description: "DeepSeek V3.2 on OllamaCloud"
+      },
+      {
+        id: "oc@gemini-3-pro-preview",
+        name: "Gemini 3 Pro Preview",
+        context: "1M",
+        description: "Gemini 3 Pro on OllamaCloud"
+      },
+      {
+        id: "oc@kimi-k2.5",
+        name: "Kimi K2.5",
+        context: "262K",
+        description: "Kimi K2.5 on OllamaCloud"
+      },
+      {
+        id: "oc@qwen3-coder-next",
+        name: "Qwen3 Coder Next",
+        context: "262K",
+        description: "Qwen3 Coder on OllamaCloud"
+      },
+      {
+        id: "oc@minimax-m2.1",
+        name: "MiniMax M2.1",
+        context: "205K",
+        description: "MiniMax M2.1 on OllamaCloud"
+      }
     ]
   };
   const providerDisplay = provider.charAt(0).toUpperCase() + provider.slice(1);
@@ -35474,23 +35683,76 @@ var init_model_selector = __esm(() => {
     ll: "LiteLLM"
   };
   ALL_PROVIDER_CHOICES = [
-    { name: "Skip (keep Claude default)", value: "skip", description: "Use native Claude model for this tier" },
+    {
+      name: "Skip (keep Claude default)",
+      value: "skip",
+      description: "Use native Claude model for this tier"
+    },
     { name: "OpenRouter", value: "openrouter", description: "580+ models via unified API" },
     { name: "OpenCode Zen", value: "zen", description: "Free models, no API key needed" },
-    { name: "Google Gemini", value: "google", description: "Direct API (GEMINI_API_KEY)", envVar: "GEMINI_API_KEY" },
-    { name: "OpenAI", value: "openai", description: "Direct API (OPENAI_API_KEY)", envVar: "OPENAI_API_KEY" },
-    { name: "xAI / Grok", value: "xai", description: "Direct API (XAI_API_KEY)", envVar: "XAI_API_KEY" },
-    { name: "MiniMax", value: "minimax", description: "Direct API (MINIMAX_API_KEY)", envVar: "MINIMAX_API_KEY" },
-    { name: "MiniMax Coding", value: "minimax-coding", description: "MiniMax Coding subscription (MINIMAX_CODING_API_KEY)", envVar: "MINIMAX_CODING_API_KEY" },
-    { name: "Kimi / Moonshot", value: "kimi", description: "Direct API (MOONSHOT_API_KEY)", envVar: "MOONSHOT_API_KEY" },
-    { name: "Kimi Coding", value: "kimi-coding", description: "Kimi Coding subscription (KIMI_CODING_API_KEY)", envVar: "KIMI_CODING_API_KEY" },
-    { name: "GLM / Zhipu", value: "glm", description: "Direct API (ZHIPU_API_KEY)", envVar: "ZHIPU_API_KEY" },
-    { name: "GLM Coding Plan", value: "glm-coding", description: "GLM Coding subscription (GLM_CODING_API_KEY)", envVar: "GLM_CODING_API_KEY" },
+    {
+      name: "Google Gemini",
+      value: "google",
+      description: "Direct API (GEMINI_API_KEY)",
+      envVar: "GEMINI_API_KEY"
+    },
+    {
+      name: "OpenAI",
+      value: "openai",
+      description: "Direct API (OPENAI_API_KEY)",
+      envVar: "OPENAI_API_KEY"
+    },
+    {
+      name: "xAI / Grok",
+      value: "xai",
+      description: "Direct API (XAI_API_KEY)",
+      envVar: "XAI_API_KEY"
+    },
+    {
+      name: "MiniMax",
+      value: "minimax",
+      description: "Direct API (MINIMAX_API_KEY)",
+      envVar: "MINIMAX_API_KEY"
+    },
+    {
+      name: "MiniMax Coding",
+      value: "minimax-coding",
+      description: "MiniMax Coding subscription (MINIMAX_CODING_API_KEY)",
+      envVar: "MINIMAX_CODING_API_KEY"
+    },
+    {
+      name: "Kimi / Moonshot",
+      value: "kimi",
+      description: "Direct API (MOONSHOT_API_KEY)",
+      envVar: "MOONSHOT_API_KEY"
+    },
+    {
+      name: "Kimi Coding",
+      value: "kimi-coding",
+      description: "Kimi Coding subscription (KIMI_CODING_API_KEY)",
+      envVar: "KIMI_CODING_API_KEY"
+    },
+    {
+      name: "GLM / Zhipu",
+      value: "glm",
+      description: "Direct API (ZHIPU_API_KEY)",
+      envVar: "ZHIPU_API_KEY"
+    },
+    {
+      name: "GLM Coding Plan",
+      value: "glm-coding",
+      description: "GLM Coding subscription (GLM_CODING_API_KEY)",
+      envVar: "GLM_CODING_API_KEY"
+    },
     { name: "Z.AI", value: "zai", description: "Z.AI API (ZAI_API_KEY)", envVar: "ZAI_API_KEY" },
     { name: "OllamaCloud", value: "ollamacloud", description: "Cloud models (OLLAMA_API_KEY)" },
     { name: "Ollama (local)", value: "ollama", description: "Local Ollama instance" },
     { name: "LM Studio (local)", value: "lmstudio", description: "Local LM Studio instance" },
-    { name: "Enter custom model", value: "custom", description: "Type a provider@model specification" }
+    {
+      name: "Enter custom model",
+      value: "custom",
+      description: "Type a provider@model specification"
+    }
   ];
   PROVIDER_MODEL_PREFIX = {
     google: "google@",
@@ -41347,10 +41609,12 @@ var init_openai_adapter = __esm(() => {
           result.push({
             type: "message",
             role: msg.role,
-            content: [{
-              type: msg.role === "user" ? "input_text" : "output_text",
-              text: msg.content
-            }]
+            content: [
+              {
+                type: msg.role === "user" ? "input_text" : "output_text",
+                text: msg.content
+              }
+            ]
           });
           continue;
         }
@@ -62090,7 +62354,7 @@ class LocalTransport {
     this.config = config3;
     this.modelName = modelName;
     this.name = config3.name;
-    this.displayName = DISPLAY_NAMES[config3.name] || "Local";
+    this.displayName = DISPLAY_NAMES2[config3.name] || "Local";
     this.concurrency = options?.concurrency;
     const envContextWindow = process.env.CLAUDISH_CONTEXT_WINDOW;
     if (envContextWindow) {
@@ -62271,7 +62535,7 @@ class LocalTransport {
     }
   }
 }
-var localProviderAgent, DISPLAY_NAMES;
+var localProviderAgent, DISPLAY_NAMES2;
 var init_local = __esm(() => {
   init_local_queue();
   init_logger();
@@ -62282,7 +62546,7 @@ var init_local = __esm(() => {
     keepAliveTimeout: 30000,
     keepAliveMaxTimeout: 600000
   });
-  DISPLAY_NAMES = {
+  DISPLAY_NAMES2 = {
     ollama: "Ollama",
     lmstudio: "LM Studio",
     vllm: "vLLM",
@@ -63973,7 +64237,10 @@ data: ${JSON.stringify(data)}
                   }
                   if (part.text) {
                     if (thinkingStarted) {
-                      send("content_block_stop", { type: "content_block_stop", index: thinkingIdx });
+                      send("content_block_stop", {
+                        type: "content_block_stop",
+                        index: thinkingIdx
+                      });
                       thinkingStarted = false;
                     }
                     let cleanedText = part.text;
@@ -64003,7 +64270,10 @@ data: ${JSON.stringify(data)}
                   }
                   if (part.functionCall) {
                     if (thinkingStarted) {
-                      send("content_block_stop", { type: "content_block_stop", index: thinkingIdx });
+                      send("content_block_stop", {
+                        type: "content_block_stop",
+                        index: thinkingIdx
+                      });
                       thinkingStarted = false;
                     }
                     if (textStarted) {
@@ -64935,7 +65205,10 @@ var init_anthropic_passthrough_adapter = __esm(() => {
       const filteredContent = message.content.map((block) => {
         if (block.type === "tool_result" && Array.isArray(block.content)) {
           const filtered = block.content.filter((c) => c.type !== "tool_reference");
-          return { ...block, content: filtered.length > 0 ? filtered : [{ type: "text", text: "" }] };
+          return {
+            ...block,
+            content: filtered.length > 0 ? filtered : [{ type: "text", text: "" }]
+          };
         }
         return block;
       }).filter((block) => block.type !== "tool_reference");
@@ -65242,6 +65515,117 @@ var init_pricing_cache = __esm(() => {
   };
 });
 
+// src/handlers/fallback-handler.ts
+class FallbackHandler {
+  candidates;
+  constructor(candidates) {
+    this.candidates = candidates;
+  }
+  async handle(c, payload) {
+    const errors6 = [];
+    for (let i = 0;i < this.candidates.length; i++) {
+      const { name, handler } = this.candidates[i];
+      const isLast = i === this.candidates.length - 1;
+      try {
+        const response = await handler.handle(c, payload);
+        if (response.ok) {
+          if (errors6.length > 0) {
+            logStderr(`[Fallback] ${name} succeeded after ${errors6.length} failed attempt(s)`);
+          }
+          return response;
+        }
+        const errorBody = await response.clone().text();
+        if (!isRetryableError(response.status, errorBody)) {
+          if (errors6.length > 0) {
+            errors6.push({ provider: name, status: response.status, message: errorBody });
+            return this.formatCombinedError(c, errors6, payload.model);
+          }
+          return response;
+        }
+        errors6.push({ provider: name, status: response.status, message: errorBody });
+        if (!isLast) {
+          logStderr(`[Fallback] ${name} failed (HTTP ${response.status}), trying next provider...`);
+        }
+      } catch (err) {
+        errors6.push({ provider: name, status: 0, message: err.message });
+        if (!isLast) {
+          logStderr(`[Fallback] ${name} error: ${err.message}, trying next provider...`);
+        }
+      }
+    }
+    return this.formatCombinedError(c, errors6, payload.model);
+  }
+  formatCombinedError(c, errors6, modelName) {
+    const summary = errors6.map((e) => `  ${e.provider}: HTTP ${e.status || "ERR"} — ${truncate(parseErrorMessage(e.message), 150)}`).join(`
+`);
+    logStderr(`[Fallback] All ${errors6.length} provider(s) failed for ${modelName || "model"}:
+${summary}`);
+    return c.json({
+      error: {
+        type: "all_providers_failed",
+        message: `All ${errors6.length} providers failed for model '${modelName || "unknown"}'`,
+        attempts: errors6.map((e) => ({
+          provider: e.provider,
+          status: e.status,
+          error: truncate(parseErrorMessage(e.message), 200)
+        }))
+      }
+    }, 502);
+  }
+  async shutdown() {
+    for (const { handler } of this.candidates) {
+      if (typeof handler.shutdown === "function") {
+        await handler.shutdown();
+      }
+    }
+  }
+}
+function isRetryableError(status, errorBody) {
+  if (status === 401 || status === 403)
+    return true;
+  if (status === 402)
+    return true;
+  if (status === 404)
+    return true;
+  if (status === 429)
+    return true;
+  const lower = errorBody.toLowerCase();
+  if (status === 422) {
+    if (lower.includes("not available") || lower.includes("model not found") || lower.includes("not supported")) {
+      return true;
+    }
+  }
+  if (status === 400) {
+    if (lower.includes("model not found") || lower.includes("not registered") || lower.includes("does not exist") || lower.includes("unknown model") || lower.includes("unsupported model")) {
+      return true;
+    }
+  }
+  if (status === 500) {
+    if (lower.includes("insufficient balance") || lower.includes("insufficient credit") || lower.includes("quota exceeded") || lower.includes("billing")) {
+      return true;
+    }
+  }
+  return false;
+}
+function parseErrorMessage(body) {
+  try {
+    const parsed = JSON.parse(body);
+    if (typeof parsed.error === "string")
+      return parsed.error;
+    if (typeof parsed.error?.message === "string")
+      return parsed.error.message;
+    if (typeof parsed.message === "string")
+      return parsed.message;
+  } catch {}
+  return body;
+}
+function truncate(s, max) {
+  return s.length > max ? s.slice(0, max) + "..." : s;
+}
+var init_fallback_handler = __esm(() => {
+  init_logger();
+});
+
 // src/proxy-server.ts
 var exports_proxy_server = {};
 __export(exports_proxy_server, {
@@ -65259,7 +65643,10 @@ async function createProxyServer(port, openrouterApiKey, model, monitorMode = fa
     if (!openRouterHandlers.has(modelId)) {
       const orProvider = new OpenRouterProvider(openrouterApiKey || "");
       const orAdapter = new OpenRouterAdapter(modelId);
-      openRouterHandlers.set(modelId, new ComposedHandler(orProvider, modelId, modelId, port, { adapter: orAdapter, isInteractive: options.isInteractive }));
+      openRouterHandlers.set(modelId, new ComposedHandler(orProvider, modelId, modelId, port, {
+        adapter: orAdapter,
+        isInteractive: options.isInteractive
+      }));
     }
     return openRouterHandlers.get(modelId);
   };
@@ -65272,7 +65659,9 @@ async function createProxyServer(port, openrouterApiKey, model, monitorMode = fa
     const modelId = targetModel.replace(/^poe:/, "");
     if (!poeHandlers.has(modelId)) {
       const poeTransport = new PoeProvider(poeApiKey);
-      poeHandlers.set(modelId, new ComposedHandler(poeTransport, modelId, modelId, port, { isInteractive: options.isInteractive }));
+      poeHandlers.set(modelId, new ComposedHandler(poeTransport, modelId, modelId, port, {
+        isInteractive: options.isInteractive
+      }));
     }
     return poeHandlers.get(modelId);
   };
@@ -65487,6 +65876,7 @@ async function createProxyServer(port, openrouterApiKey, model, monitorMode = fa
       log("[Proxy] LiteLLM model cache pre-warmed for auto-routing");
     }).catch(() => {});
   }
+  const fallbackHandlerCache = new Map;
   const getHandlerForRequest = (requestedModel) => {
     if (monitorMode)
       return nativeHandler;
@@ -65514,6 +65904,38 @@ async function createProxyServer(port, openrouterApiKey, model, monitorMode = fa
         }
       }
     }
+    {
+      const parsedForFallback = parseModelSpec(target);
+      if (!parsedForFallback.isExplicitProvider && parsedForFallback.provider !== "native-anthropic" && !isPoeModel(target)) {
+        const cacheKey2 = `fallback:${target}`;
+        if (fallbackHandlerCache.has(cacheKey2)) {
+          return fallbackHandlerCache.get(cacheKey2);
+        }
+        const chain = getFallbackChain(parsedForFallback.model, parsedForFallback.provider);
+        if (chain.length > 0) {
+          const candidates = [];
+          for (const route of chain) {
+            let handler = null;
+            if (route.provider === "openrouter") {
+              handler = getOpenRouterHandler(route.modelSpec);
+            } else {
+              handler = getRemoteProviderHandler(route.modelSpec);
+            }
+            if (handler) {
+              candidates.push({ name: route.displayName, handler });
+            }
+          }
+          if (candidates.length > 0) {
+            const resultHandler = candidates.length > 1 ? new FallbackHandler(candidates) : candidates[0].handler;
+            fallbackHandlerCache.set(cacheKey2, resultHandler);
+            if (!options.quiet && candidates.length > 1) {
+              logStderr(`[Fallback] ${candidates.length} providers for ${parsedForFallback.model}: ${candidates.map((c) => c.name).join(" → ")}`);
+            }
+            return resultHandler;
+          }
+        }
+      }
+    }
     if (isPoeModel(target)) {
       const poeHandler = getPoeHandler(target);
       if (poeHandler) {
@@ -65625,6 +66047,8 @@ var init_proxy_server = __esm(() => {
   init_pricing_cache();
   init_model_loader();
   init_model_catalog_resolver();
+  init_fallback_handler();
+  init_auto_route();
 });
 
 // src/index.ts

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claudish",
-  "version": "5.7.1",
+  "version": "5.9.0",
   "description": "Run Claude Code with any model - OpenRouter, Ollama, LM Studio & local models",
   "type": "module",
   "main": "./dist/index.js",
@@ -18,7 +18,8 @@
     "typecheck": "tsc --noEmit",
     "lint": "biome check .",
     "format": "biome format --write .",
-    "test": "bun test"
+    "test": "bun test",
+    "smoke": "bun run scripts/smoke-test.ts"
   },
   "dependencies": {
     "@inquirer/prompts": "^8.0.1",

package/recommended-models.json

@@ -1,6 +1,6 @@
 {
   "version": "1.2.0",
-  "lastUpdated": "2026-03-06",
+  "lastUpdated": "2026-03-14",
   "source": "https://openrouter.ai/models?categories=programming&fmt=cards&order=top-weekly",
   "models": [
     {
@@ -12,9 +12,9 @@
       "category": "programming",
       "priority": 1,
       "pricing": {
-        "input": "$0.29/1M",
+        "input": "$0.25/1M",
         "output": "$1.20/1M",
-        "average": "$0.75/1M"
+        "average": "$0.72/1M"
       },
       "context": "196K",
       "maxOutputTokens": 196608,
@@ -56,12 +56,12 @@
       "category": "reasoning",
       "priority": 3,
       "pricing": {
-        "input": "$0.80/1M",
-        "output": "$2.56/1M",
-        "average": "$1.68/1M"
+        "input": "$0.72/1M",
+        "output": "$2.30/1M",
+        "average": "$1.51/1M"
       },
       "context": "202K",
-      "maxOutputTokens": null,
+      "maxOutputTokens": 131072,
       "modality": "text->text",
       "supportsTools": true,
       "supportsReasoning": true,