@probeo/anymodel 0.1.0 → 0.2.0

package/README.md

@@ -49,18 +49,39 @@ for await (const chunk of stream) {
 }
 ```
 
+## Supported Providers
+
+Set the env var and go. Models are auto-discovered from each provider's API.
+
+| Provider | Env Var | Example Model |
+|----------|---------|---------------|
+| OpenAI | `OPENAI_API_KEY` | `openai/gpt-4o` |
+| Anthropic | `ANTHROPIC_API_KEY` | `anthropic/claude-sonnet-4-6` |
+| Google | `GOOGLE_API_KEY` | `google/gemini-2.5-pro` |
+| Mistral | `MISTRAL_API_KEY` | `mistral/mistral-large-latest` |
+| Groq | `GROQ_API_KEY` | `groq/llama-3.3-70b-versatile` |
+| DeepSeek | `DEEPSEEK_API_KEY` | `deepseek/deepseek-chat` |
+| xAI | `XAI_API_KEY` | `xai/grok-3` |
+| Together | `TOGETHER_API_KEY` | `together/meta-llama/Llama-3.3-70B-Instruct-Turbo` |
+| Fireworks | `FIREWORKS_API_KEY` | `fireworks/accounts/fireworks/models/llama-v3p3-70b-instruct` |
+| Perplexity | `PERPLEXITY_API_KEY` | `perplexity/sonar-pro` |
+| Ollama | `OLLAMA_BASE_URL` | `ollama/llama3.3` |
+
+Ollama runs locally with no API key — just set `OLLAMA_BASE_URL` (defaults to `http://localhost:11434/v1`).
+
 ## Model Naming
 
 Models use `provider/model` format:
 
 ```
 anthropic/claude-sonnet-4-6
-anthropic/claude-opus-4-6
-anthropic/claude-haiku-4-5
 openai/gpt-4o
-openai/o3
 google/gemini-2.5-pro
-google/gemini-2.5-flash
+mistral/mistral-large-latest
+groq/llama-3.3-70b-versatile
+deepseek/deepseek-chat
+xai/grok-3
+ollama/llama3.3
 ```
 
 ## Fallback Routing

package/dist/cli.cjs

@@ -623,7 +623,19 @@ function createOpenAIAdapter(apiKey, baseURL) {
     async listModels() {
       const res = await makeRequest("/models", void 0, "GET");
       const data = await res.json();
-      return (data.data || []).filter((m) => m.id.startsWith("gpt-") || m.id.startsWith("o") || m.id.startsWith("chatgpt-")).map((m) => ({
+      return (data.data || []).filter((m) => {
+        const id = m.id;
+        if (id.includes("embedding")) return false;
+        if (id.includes("whisper")) return false;
+        if (id.includes("tts")) return false;
+        if (id.includes("dall-e")) return false;
+        if (id.includes("davinci")) return false;
+        if (id.includes("babbage")) return false;
+        if (id.includes("moderation")) return false;
+        if (id.includes("realtime")) return false;
+        if (id.startsWith("ft:")) return false;
+        return id.startsWith("gpt-") || id.startsWith("o1") || id.startsWith("o3") || id.startsWith("o4") || id.startsWith("chatgpt-");
+      }).map((m) => ({
         id: `openai/${m.id}`,
         name: m.id,
         created: m.created,
@@ -689,10 +701,16 @@ var SUPPORTED_PARAMS2 = /* @__PURE__ */ new Set([
   "tool_choice",
   "response_format"
 ]);
-var KNOWN_MODELS = [
-  { id: "anthropic/claude-opus-4-6", name: "Claude Opus 4.6", created: 0, description: "Most capable model", context_length: 2e5, pricing: { prompt: "0.000005", completion: "0.000025" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 32768, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
+var FALLBACK_MODELS = [
+  // Claude 4.6
+  { id: "anthropic/claude-opus-4-6", name: "Claude Opus 4.6", created: 0, description: "Most capable model", context_length: 2e5, pricing: { prompt: "0.000015", completion: "0.000075" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 32768, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
   { id: "anthropic/claude-sonnet-4-6", name: "Claude Sonnet 4.6", created: 0, description: "Best balance of speed and capability", context_length: 2e5, pricing: { prompt: "0.000003", completion: "0.000015" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 16384, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
-  { id: "anthropic/claude-haiku-4-5", name: "Claude Haiku 4.5", created: 0, description: "Fastest and most compact", context_length: 2e5, pricing: { prompt: "0.000001", completion: "0.000005" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) }
+  // Claude 4.5
+  { id: "anthropic/claude-sonnet-4-5-20251022", name: "Claude Sonnet 4.5", created: 0, description: "Previous generation balanced model", context_length: 2e5, pricing: { prompt: "0.000003", completion: "0.000015" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 16384, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
+  { id: "anthropic/claude-haiku-4-5", name: "Claude Haiku 4.5", created: 0, description: "Fast and compact", context_length: 2e5, pricing: { prompt: "0.000001", completion: "0.000005" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
+  // Claude 3.5
+  { id: "anthropic/claude-3-5-sonnet-20241022", name: "Claude 3.5 Sonnet", created: 0, description: "Legacy balanced model", context_length: 2e5, pricing: { prompt: "0.000003", completion: "0.000015" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) },
+  { id: "anthropic/claude-3-5-haiku-20241022", name: "Claude 3.5 Haiku", created: 0, description: "Legacy fast model", context_length: 2e5, pricing: { prompt: "0.0000008", completion: "0.000004" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image"], output_modalities: ["text"], tokenizer: "claude" }, top_provider: { context_length: 2e5, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS2) }
 ];
 function createAnthropicAdapter(apiKey) {
   async function makeRequest(path, body, stream = false) {
@@ -952,7 +970,40 @@ ${body.system}` : jsonInstruction;
       };
     },
     async listModels() {
-      return KNOWN_MODELS;
+      try {
+        const res = await fetch(`${ANTHROPIC_API_BASE}/models`, {
+          method: "GET",
+          headers: {
+            "x-api-key": apiKey,
+            "anthropic-version": ANTHROPIC_VERSION
+          }
+        });
+        if (!res.ok) return FALLBACK_MODELS;
+        const data = await res.json();
+        const models = data.data || [];
+        return models.filter((m) => m.type === "model").map((m) => ({
+          id: `anthropic/${m.id}`,
+          name: m.display_name || m.id,
+          created: m.created_at ? new Date(m.created_at).getTime() / 1e3 : 0,
+          description: m.display_name || "",
+          context_length: 2e5,
+          pricing: { prompt: "0", completion: "0" },
+          architecture: {
+            modality: "text+image->text",
+            input_modalities: ["text", "image"],
+            output_modalities: ["text"],
+            tokenizer: "claude"
+          },
+          top_provider: {
+            context_length: 2e5,
+            max_completion_tokens: 16384,
+            is_moderated: false
+          },
+          supported_parameters: Array.from(SUPPORTED_PARAMS2)
+        }));
+      } catch {
+        return FALLBACK_MODELS;
+      }
     },
     supportsParameter(param) {
       return SUPPORTED_PARAMS2.has(param);
@@ -990,9 +1041,16 @@ var SUPPORTED_PARAMS3 = /* @__PURE__ */ new Set([
   "tool_choice",
   "response_format"
 ]);
-var KNOWN_MODELS2 = [
+var FALLBACK_MODELS2 = [
+  // Gemini 2.5
   { id: "google/gemini-2.5-pro", name: "Gemini 2.5 Pro", created: 0, description: "Most capable Gemini model", context_length: 1048576, pricing: { prompt: "0.00000125", completion: "0.000005" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 65536, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) },
-  { id: "google/gemini-2.5-flash", name: "Gemini 2.5 Flash", created: 0, description: "Fast and efficient", context_length: 1048576, pricing: { prompt: "0.00000015", completion: "0.0000006" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 65536, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) }
+  { id: "google/gemini-2.5-flash", name: "Gemini 2.5 Flash", created: 0, description: "Fast and efficient", context_length: 1048576, pricing: { prompt: "0.00000015", completion: "0.0000006" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 65536, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) },
+  // Gemini 2.0
+  { id: "google/gemini-2.0-flash", name: "Gemini 2.0 Flash", created: 0, description: "Fast multimodal model", context_length: 1048576, pricing: { prompt: "0.0000001", completion: "0.0000004" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 65536, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) },
+  { id: "google/gemini-2.0-flash-lite", name: "Gemini 2.0 Flash Lite", created: 0, description: "Lightweight and fast", context_length: 1048576, pricing: { prompt: "0.00000005", completion: "0.0000002" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 65536, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) },
+  // Gemini 1.5
+  { id: "google/gemini-1.5-pro", name: "Gemini 1.5 Pro", created: 0, description: "Previous generation pro model", context_length: 2097152, pricing: { prompt: "0.00000125", completion: "0.000005" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 2097152, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) },
+  { id: "google/gemini-1.5-flash", name: "Gemini 1.5 Flash", created: 0, description: "Previous generation flash model", context_length: 1048576, pricing: { prompt: "0.000000075", completion: "0.0000003" }, architecture: { modality: "text+image->text", input_modalities: ["text", "image", "video", "audio"], output_modalities: ["text"], tokenizer: "gemini" }, top_provider: { context_length: 1048576, max_completion_tokens: 8192, is_moderated: false }, supported_parameters: Array.from(SUPPORTED_PARAMS3) }
 ];
 function createGoogleAdapter(apiKey) {
   function getModelEndpoint(model, stream) {
@@ -1194,7 +1252,37 @@ function createGoogleAdapter(apiKey) {
       };
     },
     async listModels() {
-      return KNOWN_MODELS2;
+      try {
+        const res = await fetch(`${GEMINI_API_BASE}/models?key=${apiKey}`);
+        if (!res.ok) return FALLBACK_MODELS2;
+        const data = await res.json();
+        const models = data.models || [];
+        return models.filter((m) => m.name?.startsWith("models/gemini-") && m.supportedGenerationMethods?.includes("generateContent")).map((m) => {
+          const modelId = m.name.replace("models/", "");
+          return {
+            id: `google/${modelId}`,
+            name: m.displayName || modelId,
+            created: 0,
+            description: m.description || "",
+            context_length: m.inputTokenLimit || 1048576,
+            pricing: { prompt: "0", completion: "0" },
+            architecture: {
+              modality: "text+image->text",
+              input_modalities: ["text", "image", "video", "audio"],
+              output_modalities: ["text"],
+              tokenizer: "gemini"
+            },
+            top_provider: {
+              context_length: m.inputTokenLimit || 1048576,
+              max_completion_tokens: m.outputTokenLimit || 65536,
+              is_moderated: false
+            },
+            supported_parameters: Array.from(SUPPORTED_PARAMS3)
+          };
+        });
+      } catch {
+        return FALLBACK_MODELS2;
+      }
     },
     supportsParameter(param) {
       return SUPPORTED_PARAMS3.has(param);
@@ -1358,14 +1446,22 @@ function deepMerge(target, source) {
 }
 function envConfig() {
   const config = {};
-  if (process.env.ANTHROPIC_API_KEY) {
-    config.anthropic = { apiKey: process.env.ANTHROPIC_API_KEY };
-  }
-  if (process.env.OPENAI_API_KEY) {
-    config.openai = { apiKey: process.env.OPENAI_API_KEY };
-  }
-  if (process.env.GOOGLE_API_KEY) {
-    config.google = { apiKey: process.env.GOOGLE_API_KEY };
+  const envMap = [
+    ["openai", "OPENAI_API_KEY"],
+    ["anthropic", "ANTHROPIC_API_KEY"],
+    ["google", "GOOGLE_API_KEY"],
+    ["mistral", "MISTRAL_API_KEY"],
+    ["groq", "GROQ_API_KEY"],
+    ["deepseek", "DEEPSEEK_API_KEY"],
+    ["xai", "XAI_API_KEY"],
+    ["together", "TOGETHER_API_KEY"],
+    ["fireworks", "FIREWORKS_API_KEY"],
+    ["perplexity", "PERPLEXITY_API_KEY"]
+  ];
+  for (const [key, envVar] of envMap) {
+    if (process.env[envVar]) {
+      config[key] = { apiKey: process.env[envVar] };
+    }
   }
   return config;
 }
@@ -1780,21 +1876,42 @@ var AnyModel = class {
     };
   }
   registerProviders() {
-    const { anthropic, openai, google } = this.config;
-    const anthropicKey = anthropic?.apiKey || process.env.ANTHROPIC_API_KEY;
-    const openaiKey = openai?.apiKey || process.env.OPENAI_API_KEY;
-    const googleKey = google?.apiKey || process.env.GOOGLE_API_KEY;
+    const config = this.config;
+    const openaiKey = config.openai?.apiKey || process.env.OPENAI_API_KEY;
     if (openaiKey) {
       this.registry.register("openai", createOpenAIAdapter(openaiKey));
     }
+    const anthropicKey = config.anthropic?.apiKey || process.env.ANTHROPIC_API_KEY;
     if (anthropicKey) {
       this.registry.register("anthropic", createAnthropicAdapter(anthropicKey));
     }
+    const googleKey = config.google?.apiKey || process.env.GOOGLE_API_KEY;
     if (googleKey) {
       this.registry.register("google", createGoogleAdapter(googleKey));
     }
-    if (this.config.custom) {
-      for (const [name, customConfig] of Object.entries(this.config.custom)) {
+    const builtinProviders = [
+      { name: "mistral", baseURL: "https://api.mistral.ai/v1", configKey: "mistral", envVar: "MISTRAL_API_KEY" },
+      { name: "groq", baseURL: "https://api.groq.com/openai/v1", configKey: "groq", envVar: "GROQ_API_KEY" },
+      { name: "deepseek", baseURL: "https://api.deepseek.com", configKey: "deepseek", envVar: "DEEPSEEK_API_KEY" },
+      { name: "xai", baseURL: "https://api.x.ai/v1", configKey: "xai", envVar: "XAI_API_KEY" },
+      { name: "together", baseURL: "https://api.together.xyz/v1", configKey: "together", envVar: "TOGETHER_API_KEY" },
+      { name: "fireworks", baseURL: "https://api.fireworks.ai/inference/v1", configKey: "fireworks", envVar: "FIREWORKS_API_KEY" },
+      { name: "perplexity", baseURL: "https://api.perplexity.ai", configKey: "perplexity", envVar: "PERPLEXITY_API_KEY" }
+    ];
+    for (const { name, baseURL, configKey, envVar } of builtinProviders) {
+      const providerConfig = config[configKey];
+      const key = providerConfig?.apiKey || process.env[envVar];
+      if (key) {
+        this.registry.register(name, createCustomAdapter(name, { baseURL, apiKey: key }));
+      }
+    }
+    const ollamaConfig = config.ollama;
+    const ollamaURL = ollamaConfig?.baseURL || process.env.OLLAMA_BASE_URL || "http://localhost:11434/v1";
+    if (ollamaConfig || process.env.OLLAMA_BASE_URL) {
+      this.registry.register("ollama", createCustomAdapter("ollama", { baseURL: ollamaURL }));
+    }
+    if (config.custom) {
+      for (const [name, customConfig] of Object.entries(config.custom)) {
         this.registry.register(name, createCustomAdapter(name, customConfig));
       }
     }