osagent 0.1.26 → 0.1.27

package/dist/cli.js

@@ -53251,14 +53251,14 @@ var init_node = __esm({
       Mode2["MODE_UNSPECIFIED"] = "MODE_UNSPECIFIED";
       Mode2["MODE_DYNAMIC"] = "MODE_DYNAMIC";
     })(Mode || (Mode = {}));
-    (function(AuthType3) {
-      AuthType3["AUTH_TYPE_UNSPECIFIED"] = "AUTH_TYPE_UNSPECIFIED";
-      AuthType3["NO_AUTH"] = "NO_AUTH";
-      AuthType3["API_KEY_AUTH"] = "API_KEY_AUTH";
-      AuthType3["HTTP_BASIC_AUTH"] = "HTTP_BASIC_AUTH";
-      AuthType3["GOOGLE_SERVICE_ACCOUNT_AUTH"] = "GOOGLE_SERVICE_ACCOUNT_AUTH";
-      AuthType3["OAUTH"] = "OAUTH";
-      AuthType3["OIDC_AUTH"] = "OIDC_AUTH";
+    (function(AuthType4) {
+      AuthType4["AUTH_TYPE_UNSPECIFIED"] = "AUTH_TYPE_UNSPECIFIED";
+      AuthType4["NO_AUTH"] = "NO_AUTH";
+      AuthType4["API_KEY_AUTH"] = "API_KEY_AUTH";
+      AuthType4["HTTP_BASIC_AUTH"] = "HTTP_BASIC_AUTH";
+      AuthType4["GOOGLE_SERVICE_ACCOUNT_AUTH"] = "GOOGLE_SERVICE_ACCOUNT_AUTH";
+      AuthType4["OAUTH"] = "OAUTH";
+      AuthType4["OIDC_AUTH"] = "OIDC_AUTH";
     })(AuthType || (AuthType = {}));
     (function(ApiSpec2) {
       ApiSpec2["API_SPEC_UNSPECIFIED"] = "API_SPEC_UNSPECIFIED";
@@ -146147,27 +146147,31 @@ function createContentGeneratorConfig(config, authType, generationConfig) {
     return {
       ...newContentGeneratorConfig,
       model: newContentGeneratorConfig?.model || DEFAULT_OLLAMA_CODER_MODEL,
-      baseUrl: newContentGeneratorConfig?.baseUrl || "https://api.ollama.com",
-      // API key from OLLAMA_API_KEY env var or settings
-      apiKey: newContentGeneratorConfig?.apiKey || process.env["OLLAMA_API_KEY"]
+      // Force Ollama Cloud endpoint explicitly
+      baseUrl: "https://api.ollama.com",
+      apiKey: process.env["OLLAMA_API_KEY"]
     };
   }
   if (authType === AuthType2.OLLAMA_LOCAL) {
     return {
       ...newContentGeneratorConfig,
       model: newContentGeneratorConfig?.model || DEFAULT_OLLAMA_CODER_MODEL,
-      baseUrl: newContentGeneratorConfig?.baseUrl || "http://localhost:11434/v1",
+      // Force localhost endpoint explicitly
+      baseUrl: "http://localhost:11434/v1",
       // Local Ollama doesn't require API key
-      apiKey: newContentGeneratorConfig?.apiKey || "ollama"
+      apiKey: "ollama"
     };
   }
   if (authType === AuthType2.USE_GROQ) {
+    const groqApiKey = process.env["GROQ_API_KEY"];
     return {
       ...newContentGeneratorConfig,
       model: newContentGeneratorConfig?.model || "moonshotai/kimi-k2-instruct-0905",
-      baseUrl: newContentGeneratorConfig?.baseUrl || "https://api.groq.com/openai/v1",
-      // API key from GROQ_API_KEY env var or settings
-      apiKey: newContentGeneratorConfig?.apiKey || process.env["GROQ_API_KEY"]
+      // Force GROQ endpoint explicitly - this overrides any inherited baseUrl
+      baseUrl: "https://api.groq.com/openai/v1",
+      apiKey: groqApiKey,
+      // Disable cache control for GROQ (not supported)
+      disableCacheControl: true
     };
   }
   return {
@@ -146176,7 +146180,7 @@ function createContentGeneratorConfig(config, authType, generationConfig) {
   };
 }
 async function createContentGenerator(config, gcConfig, sessionId2, isInitialAuth) {
-  const version2 = "0.1.26";
+  const version2 = "0.1.27";
   const userAgent2 = `OSAgent/${version2} (${process.platform}; ${process.arch})`;
   const baseHeaders = {
     "User-Agent": userAgent2
@@ -146251,16 +146255,16 @@ var init_contentGenerator = __esm({
     init_models();
     init_installationManager();
     init_loggingContentGenerator();
-    (function(AuthType3) {
-      AuthType3["LOGIN_WITH_OSAGENT"] = "oauth-personal";
-      AuthType3["USE_OSA"] = "OSA-api-key";
-      AuthType3["USE_VERTEX_AI"] = "vertex-ai";
-      AuthType3["CLOUD_SHELL"] = "cloud-shell";
-      AuthType3["USE_OPENAI"] = "openai";
-      AuthType3["OSA_OAUTH"] = "OSA-oauth";
-      AuthType3["OLLAMA_CLOUD"] = "ollama-cloud";
-      AuthType3["OLLAMA_LOCAL"] = "ollama-local";
-      AuthType3["USE_GROQ"] = "groq";
+    (function(AuthType4) {
+      AuthType4["LOGIN_WITH_OSAGENT"] = "oauth-personal";
+      AuthType4["USE_OSA"] = "OSA-api-key";
+      AuthType4["USE_VERTEX_AI"] = "vertex-ai";
+      AuthType4["CLOUD_SHELL"] = "cloud-shell";
+      AuthType4["USE_OPENAI"] = "openai";
+      AuthType4["OSA_OAUTH"] = "OSA-oauth";
+      AuthType4["OLLAMA_CLOUD"] = "ollama-cloud";
+      AuthType4["OLLAMA_LOCAL"] = "ollama-local";
+      AuthType4["USE_GROQ"] = "groq";
     })(AuthType2 || (AuthType2 = {}));
     __name(createContentGeneratorConfig, "createContentGeneratorConfig");
     __name(createContentGenerator, "createContentGenerator");
@@ -309775,7 +309779,7 @@ __name(getPackageJson, "getPackageJson");
 // packages/cli/src/utils/version.ts
 async function getCliVersion() {
   const pkgJson = await getPackageJson();
-  return "0.1.26";
+  return "0.1.27";
 }
 __name(getCliVersion, "getCliVersion");
 
@@ -313944,8 +313948,8 @@ var formatDuration = /* @__PURE__ */ __name((milliseconds) => {
 
 // packages/cli/src/generated/git-commit.ts
 init_esbuild_shims();
-var GIT_COMMIT_INFO2 = "20c4873";
-var CLI_VERSION2 = "0.1.26";
+var GIT_COMMIT_INFO2 = "fbe9649";
+var CLI_VERSION2 = "0.1.27";
 
 // packages/cli/src/utils/systemInfo.ts
 async function getNpmVersion() {
@@ -317601,7 +317605,10 @@ var AVAILABLE_MODELS_OSA = [
       return t2(
         "Qwen3-Coder 480B Cloud - Most powerful agentic coding model with 256K context (Recommended)"
       );
-    }
+    },
+    contextWindow: 262144,
+    // 256K
+    tokensPerSecond: 100
   },
   {
     id: "qwen3-coder:30b",
@@ -317610,7 +317617,10 @@ var AVAILABLE_MODELS_OSA = [
       return t2(
         "Qwen3-Coder 30B - Efficient coding model with 30B params, only 3.3B activated"
       );
-    }
+    },
+    contextWindow: 131072,
+    // 131K
+    tokensPerSecond: 150
   },
   {
     id: MAINLINE_VLM,
@@ -317620,7 +317630,10 @@ var AVAILABLE_MODELS_OSA = [
         "Vision model for multimodal tasks"
       );
     },
-    isVision: true
+    isVision: true,
+    contextWindow: 131072,
+    // 131K
+    tokensPerSecond: 50
   }
 ];
 var AVAILABLE_MODELS_GROQ = [
@@ -317631,37 +317644,42 @@ var AVAILABLE_MODELS_GROQ = [
       return t2(
         "Kimi K2 0905 - Best coding model with 256K context, ~200 tok/s on GROQ"
       );
-    }
+    },
+    contextWindow: 262144,
+    // 256K
+    tokensPerSecond: 200
   },
   {
     id: "moonshotai/kimi-k2-instruct",
-    label: "Kimi K2 Instruct",
+    label: "Kimi K2 Instruct (Legacy)",
     get description() {
       return t2(
-        "Kimi K2 Instruct - 131K context, excellent for coding and tool use"
+        "Kimi K2 Instruct - 131K context (deprecated, use 0905 instead)"
       );
-    }
+    },
+    contextWindow: 131072,
+    // 131K
+    tokensPerSecond: 200
   },
   {
     id: "llama-3.3-70b-versatile",
     label: "Llama 3.3 70B",
     get description() {
       return t2("Llama 3.3 70B - Versatile model for general coding tasks");
-    }
+    },
+    contextWindow: 131072,
+    // 131K
+    tokensPerSecond: 280
   },
   {
     id: "llama-3.1-8b-instant",
     label: "Llama 3.1 8B Instant",
     get description() {
       return t2("Llama 3.1 8B - Ultra-fast for simple tasks");
-    }
-  },
-  {
-    id: "mixtral-8x7b-32768",
-    label: "Mixtral 8x7B",
-    get description() {
-      return t2("Mixtral 8x7B MoE - 32K context, good balance of speed/quality");
-    }
+    },
+    contextWindow: 131072,
+    // 131K
+    tokensPerSecond: 560
   }
 ];
 var AVAILABLE_MODELS_OLLAMA_LOCAL = [
@@ -317671,7 +317689,11 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("Qwen2.5-Coder 32B - Best local coding model (requires ~20GB VRAM)");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 131072,
+    // 128K
+    tokensPerSecond: 30
+    // Varies by hardware
   },
   {
     id: "qwen2.5-coder:14b",
@@ -317679,7 +317701,10 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("Qwen2.5-Coder 14B - Great local coding model (requires ~10GB VRAM)");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 131072,
+    // 128K
+    tokensPerSecond: 50
   },
   {
     id: "qwen2.5-coder:7b",
@@ -317687,7 +317712,10 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("Qwen2.5-Coder 7B - Good local coding model (requires ~5GB VRAM)");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 131072,
+    // 128K
+    tokensPerSecond: 80
   },
   {
     id: "codellama:34b",
@@ -317695,7 +317723,10 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("CodeLlama 34B - Meta coding model (requires ~20GB VRAM)");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 16384,
+    // 16K
+    tokensPerSecond: 25
   },
   {
     id: "deepseek-coder-v2:16b",
@@ -317703,7 +317734,10 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("DeepSeek Coder V2 16B - Efficient MoE coding model");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 131072,
+    // 128K
+    tokensPerSecond: 60
   },
   {
     id: "llama3.2:latest",
@@ -317711,7 +317745,10 @@ var AVAILABLE_MODELS_OLLAMA_LOCAL = [
     get description() {
       return t2("Llama 3.2 - General purpose model with coding capabilities");
     },
-    isLocal: true
+    isLocal: true,
+    contextWindow: 131072,
+    // 128K
+    tokensPerSecond: 70
   }
 ];
 function getOpenAIAvailableModelFromEnv() {
@@ -317842,11 +317879,8 @@ var providerCommand = {
         const apiKey = process.env["GROQ_API_KEY"];
         if (!apiKey) {
           return {
-            type: "message",
-            messageType: "error",
-            content: t2(`GROQ API key required. Set GROQ_API_KEY environment variable.
-
-Get your API key at: https://console.groq.com/keys`)
+            type: "start_auth",
+            authType: AuthType2.USE_GROQ
           };
         }
         try {
@@ -317923,9 +317957,8 @@ Use /model to select a local model.`)
         const apiKey = process.env["OPENAI_API_KEY"];
         if (!apiKey) {
           return {
-            type: "message",
-            messageType: "error",
-            content: t2(`OpenAI API key required. Set OPENAI_API_KEY environment variable.`)
+            type: "start_auth",
+            authType: AuthType2.USE_OPENAI
           };
         }
         try {
@@ -351462,6 +351495,31 @@ var DialogManager = /* @__PURE__ */ __name(({
         }
       );
     }
+    if (uiState.pendingAuthType === AuthType2.USE_GROQ) {
+      return /* @__PURE__ */ (0, import_jsx_runtime86.jsx)(
+        OpenAIKeyPrompt,
+        {
+          onSubmit: (apiKey, _baseUrl, model) => {
+            process34.env["GROQ_API_KEY"] = apiKey;
+            uiActions.handleAuthSelect(AuthType2.USE_GROQ, "User" /* User */, {
+              apiKey,
+              // Always use GROQ endpoint - ignore user-provided baseUrl
+              baseUrl: "https://api.groq.com/openai/v1",
+              model: model || "moonshotai/kimi-k2-instruct-0905"
+            });
+          },
+          onCancel: () => {
+            uiActions.cancelAuthentication();
+            uiActions.setAuthState("updating" /* Updating */);
+          },
+          defaultApiKey: process34.env["GROQ_API_KEY"] || "",
+          defaultBaseUrl: "https://api.groq.com/openai/v1",
+          defaultModel: "moonshotai/kimi-k2-instruct-0905",
+          providerName: "GROQ",
+          apiKeyUrl: "https://console.groq.com/keys"
+        }
+      );
+    }
     if (uiState.pendingAuthType === AuthType2.OSA_OAUTH) {
       return /* @__PURE__ */ (0, import_jsx_runtime86.jsx)(
         OSAOAuthProgress,
@@ -355164,6 +355222,18 @@ var useAuthCommand = /* @__PURE__ */ __name((settings, config, addItem) => {
         await performAuth(authType, scope, credentials);
         return;
       }
+      if (authType === AuthType2.USE_GROQ) {
+        if (credentials) {
+          config.updateCredentials({
+            apiKey: credentials.apiKey,
+            // Always use GROQ endpoint - ignore user-provided baseUrl
+            baseUrl: "https://api.groq.com/openai/v1",
+            model: credentials.model || "moonshotai/kimi-k2-instruct-0905"
+          });
+          await performAuth(authType, scope, credentials);
+        }
+        return;
+      }
       await performAuth(authType, scope);
     },
     [config, performAuth]
@@ -355171,6 +355241,12 @@ var useAuthCommand = /* @__PURE__ */ __name((settings, config, addItem) => {
   const openAuthDialog = (0, import_react106.useCallback)(() => {
     setIsAuthDialogOpen(true);
   }, []);
+  const startAuthForProvider = (0, import_react106.useCallback)((authType) => {
+    setPendingAuthType(authType);
+    setAuthError(null);
+    setIsAuthDialogOpen(false);
+    setIsAuthenticating(true);
+  }, []);
   const cancelAuthentication = (0, import_react106.useCallback)(() => {
     if (isAuthenticating && pendingAuthType === AuthType2.OSA_OAUTH) {
       cancelOSAAuth();
@@ -355214,6 +355290,7 @@ var useAuthCommand = /* @__PURE__ */ __name((settings, config, addItem) => {
     OSAAuthState,
     handleAuthSelect,
     openAuthDialog,
+    startAuthForProvider,
     cancelAuthentication
   };
 }, "useAuthCommand");
@@ -356100,6 +356177,10 @@ var useSlashCommandProcessor = /* @__PURE__ */ __name((config, settings, addItem
                     true
                   );
                 }
+                case "start_auth": {
+                  actions.startAuthForProvider(result.authType);
+                  return { type: "handled" };
+                }
                 default: {
                   const unhandled = result;
                   throw new Error(
@@ -362832,6 +362913,7 @@ var AppContainer = /* @__PURE__ */ __name((props) => {
     OSAAuthState,
     handleAuthSelect,
     openAuthDialog,
+    startAuthForProvider,
     cancelAuthentication
   } = useAuthCommand(settings, config, historyManager.addItem);
   const { proQuotaRequest, handleProQuotaChoice } = useQuotaAndFallback({
@@ -362921,7 +363003,8 @@ var AppContainer = /* @__PURE__ */ __name((props) => {
       addConfirmUpdateExtensionRequest,
       openSubagentCreateDialog,
       openAgentsManagerDialog,
-      _showQuitConfirmation: showQuitConfirmation
+      _showQuitConfirmation: showQuitConfirmation,
+      startAuthForProvider
     }),
     [
       openAuthDialog,
@@ -362938,7 +363021,8 @@ var AppContainer = /* @__PURE__ */ __name((props) => {
       addConfirmUpdateExtensionRequest,
       showQuitConfirmation,
       openSubagentCreateDialog,
-      openAgentsManagerDialog
+      openAgentsManagerDialog,
+      startAuthForProvider
     ]
   );
   const {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "osagent",
-  "version": "0.1.26",
+  "version": "0.1.27",
   "description": "OS Agent - AI-powered CLI for autonomous coding with Ollama Cloud and Qwen models",
   "author": "Roberto Luna",
   "license": "Apache-2.0",