tachibot-mcp 2.0.5 → 2.0.6

package/dist/src/config/model-constants.js

@@ -32,11 +32,14 @@ export const PERPLEXITY_MODELS = {
     SONAR_PRO: "sonar-pro", // Main search model
     SONAR_REASONING: "sonar-reasoning-pro", // Reasoning model
 };
-// Grok Models (xAI) - Updated 2025-11-07
+// Grok Models (xAI) - Updated 2025-11-21 with Grok 4.1
 export const GROK_MODELS = {
-    // New fast models (2025) - PRIMARY USE
+    // Grok 4.1 models (Nov 2025) - LATEST & BEST
+    _4_1: "grok-4.1", // Latest: 2M context, $0.20/$0.50, enhanced reasoning & creativity
+    _4_1_FAST: "grok-4.1-fast", // Tool-calling optimized: 2M context, $0.20/$0.50, agentic workflows
+    // Previous fast models (2025) - Still good
     CODE_FAST: "grok-code-fast-1", // Coding specialist: 256K→2M, $0.20/$1.50, 92 tok/sec
-    _4_FAST_REASONING: "grok-4-fast-reasoning", // Cheap reasoning: 2M→4M, $0.20/$0.50 (3x cheaper!)
+    _4_FAST_REASONING: "grok-4-fast-reasoning", // Cheap reasoning: 2M→4M, $0.20/$0.50
     _4_FAST: "grok-4-fast-non-reasoning", // Fast general: 2M→4M, $0.20/$0.50
     // Expensive/specialized (use sparingly)
     _4_HEAVY: "grok-4-0709", // Multi-agent: 256K→2M, $3/$15 (expensive!)
@@ -132,39 +135,39 @@ export const TOOL_DEFAULTS = {
         maxTokens: 3000,
         temperature: 0.7,
     },
-    // Grok tools - UPDATED with new fast models
+    // Grok tools - UPDATED 2025-11-21 with Grok 4.1
     grok: {
-        model: GROK_MODELS._4_FAST_REASONING, // Changed: 3x cheaper output
+        model: GROK_MODELS._4_1, // Latest: Enhanced reasoning & creativity
         maxTokens: 4000,
         temperature: 0.7,
     },
     grok_reason: {
-        model: GROK_MODELS._4_FAST_REASONING, // Changed: 3x cheaper, 8x context
+        model: GROK_MODELS._4_1, // Latest: Enhanced reasoning with lower hallucination
         maxTokens: 8000,
         temperature: 0.7,
     },
     grok_code: {
-        model: GROK_MODELS.CODE_FAST, // Changed: Coding specialist, 3x faster
+        model: GROK_MODELS._4_1_FAST, // Tool-calling optimized, better than code-fast-1
         maxTokens: 4000,
         temperature: 0.3,
     },
     grok_search: {
-        model: GROK_MODELS._4_FAST_REASONING, // Changed: Use fast reasoning with search
+        model: GROK_MODELS._4_1, // Latest with enhanced reasoning
         maxTokens: 3000,
         temperature: 0.7,
     },
     grok_brainstorm: {
-        model: GROK_MODELS._4_FAST, // Changed: Fast non-reasoning for creativity
+        model: GROK_MODELS._4_1, // Latest: Enhanced creativity & emotional intelligence
         maxTokens: 4000,
         temperature: 0.9,
     },
     grok_architect: {
-        model: GROK_MODELS._4_FAST_REASONING, // New: Architecture needs reasoning
+        model: GROK_MODELS._4_1, // Latest: Enhanced architecture reasoning
         maxTokens: 4000,
         temperature: 0.6,
     },
     grok_debug: {
-        model: GROK_MODELS.CODE_FAST, // New: Use code specialist for debugging
+        model: GROK_MODELS._4_1_FAST, // Tool-calling optimized for debugging
         maxTokens: 3000,
         temperature: 0.3,
     },

package/dist/src/tools/grok-enhanced.js

@@ -12,10 +12,13 @@ config({ path: path.resolve(__dirname, '../../../.env') });
 // Grok API configuration
 const GROK_API_KEY = process.env.GROK_API_KEY || process.env.XAI_API_KEY;
 const GROK_API_URL = "https://api.x.ai/v1/chat/completions";
-// Grok models - Updated 2025-11-07 with new fast models
+// Grok models - Updated 2025-11-21 with Grok 4.1
 export var GrokModel;
 (function (GrokModel) {
-    // New fast models (2025) - PRIMARY USE
+    // Grok 4.1 models (Nov 2025) - LATEST & BEST
+    GrokModel["GROK_4_1"] = "grok-4.1";
+    GrokModel["GROK_4_1_FAST"] = "grok-4.1-fast";
+    // Previous fast models (2025) - Still good
     GrokModel["CODE_FAST"] = "grok-code-fast-1";
     GrokModel["GROK_4_FAST_REASONING"] = "grok-4-fast-reasoning";
     GrokModel["GROK_4_FAST"] = "grok-4-fast-non-reasoning";
@@ -35,7 +38,7 @@ export async function callGrokEnhanced(messages, options = {}) {
             content: `[Grok API key not configured. Add GROK_API_KEY or XAI_API_KEY to .env file]`
         };
     }
-    const { model = GrokModel.GROK_4_FAST_REASONING, // Changed: Use cheap fast reasoning by default
+    const { model = GrokModel.GROK_4_1, // Updated 2025-11-21: Use latest Grok 4.1 by default
     temperature = 0.7, maxTokens = options.useHeavy ? 100000 : 4000, enableLiveSearch = false, searchSources = 100, // Default to 100 sources for cost control
     searchDomains = [], structuredOutput = false } = options;
     try {
@@ -140,9 +143,9 @@ Cite your sources when using web data.`
                 content: query
             }
         ];
-        log?.info(`Grok Scout: ${variant} research with ${enableLiveSearch ? 'live search' : 'knowledge base'} (using ${enableLiveSearch ? 'grok-4-fast-reasoning with search' : 'grok-4-fast-reasoning'})`);
+        log?.info(`Grok Scout: ${variant} research with ${enableLiveSearch ? 'live search' : 'knowledge base'} (using grok-4.1 with enhanced reasoning)`);
         const result = await callGrokEnhanced(messages, {
-            model: GrokModel.GROK_4_FAST_REASONING, // Changed: Use fast reasoning (works with search)
+            model: GrokModel.GROK_4_1, // Updated 2025-11-21: Use latest Grok 4.1
             enableLiveSearch,
             searchSources,
             searchDomains,
@@ -197,11 +200,11 @@ ${enableLiveSearch ? 'Use live search for current information when needed.' : ''
                 content: problem
             }
         ];
-        const modelName = useHeavy ? 'Grok-4-Heavy' : 'Grok-4-Fast-Reasoning';
-        const costInfo = useHeavy ? '$3/$15 (expensive!)' : '$0.20/$0.50 (cheap!)';
+        const modelName = useHeavy ? 'Grok-4-Heavy' : 'Grok-4.1';
+        const costInfo = useHeavy ? '$3/$15 (expensive!)' : '$0.20/$0.50 (latest!)';
         log?.info(`Using ${modelName} (${approach}) with ${enableLiveSearch ? 'live search' : 'knowledge base'} - Cost: ${costInfo}`);
         const result = await callGrokEnhanced(messages, {
-            model: useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_FAST_REASONING, // Changed: Use fast reasoning unless heavy requested
+            model: useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_1, // Updated 2025-11-21: Use latest Grok 4.1
             useHeavy,
             enableLiveSearch,
             searchSources: 50,
@@ -253,14 +256,14 @@ export const grokFunctionTool = {
         ];
         // Make request with tools
         const requestBody = {
-            model: args.useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_FAST_REASONING, // Changed: Use fast reasoning unless heavy requested
+            model: args.useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_1_FAST, // Updated 2025-11-21: Use tool-calling optimized Grok 4.1 Fast
             messages,
             tools,
             tool_choice: "auto", // Let Grok decide when to call functions
             max_tokens: args.useHeavy ? 50000 : 5000,
             temperature: 0.3 // Lower for function calling
         };
-        log?.info(`Function calling with ${args.useHeavy ? 'Grok-4-Heavy ($3/$15)' : 'Grok-4-Fast-Reasoning ($0.20/$0.50)'}`);
+        log?.info(`Function calling with ${args.useHeavy ? 'Grok-4-Heavy ($3/$15)' : 'Grok-4.1-Fast ($0.20/$0.50, tool-calling optimized)'}`);
         try {
             const response = await fetch(GROK_API_URL, {
                 method: "POST",
@@ -318,13 +321,13 @@ Limit search to ${max_search_results} sources for cost control.`
                 content: `Search for: ${query}`
             }
         ];
-        log?.info(`Grok Search: ${max_search_results} sources, recency: ${recency} (using grok-4-fast-reasoning with search)`);
+        log?.info(`Grok Search: ${max_search_results} sources, recency: ${recency} (using grok-4.1 with enhanced reasoning)`);
         // Extract domains from sources if specified
         const domains = sources
             ?.filter((s) => s.allowed_websites)
             ?.flatMap((s) => s.allowed_websites) || [];
         const result = await callGrokEnhanced(messages, {
-            model: GrokModel.GROK_4_FAST_REASONING, // Changed: Use fast reasoning with search
+            model: GrokModel.GROK_4_1, // Updated 2025-11-21: Use latest Grok 4.1 with search
             enableLiveSearch: true,
             searchSources: max_search_results,
             searchDomains: domains,
@@ -348,10 +351,10 @@ export function isGrokAvailable() {
 export function getGrokStatus() {
     return {
         available: isGrokAvailable(),
-        model: GrokModel.GROK_4_FAST_REASONING,
+        model: GrokModel.GROK_4_1,
         features: [
-            'Fast reasoning (grok-4-fast-reasoning: $0.20/$0.50, 2M→4M context)',
-            'Code specialist (grok-code-fast-1: 92 tok/sec)',
+            'Grok 4.1 (Nov 2025): Enhanced reasoning, creativity & emotional intelligence ($0.20/$0.50, 2M context)',
+            'Grok 4.1 Fast: Tool-calling optimized, agentic workflows ($0.20/$0.50, 2M context)',
             'Heavy mode available (grok-4-0709: $3/$15, use sparingly)',
             'Live web search with citations',
             'Function calling',

package/dist/src/tools/grok-tools.js

@@ -14,10 +14,13 @@ config({ path: path.resolve(__dirname, '../../../.env') });
 // Grok API configuration
 const GROK_API_KEY = process.env.GROK_API_KEY;
 const GROK_API_URL = "https://api.x.ai/v1/chat/completions";
-// Available Grok models - Updated 2025-11-07 with new fast models
+// Available Grok models - Updated 2025-11-21 with Grok 4.1
 export var GrokModel;
 (function (GrokModel) {
-    // New fast models (2025) - PRIMARY USE
+    // Grok 4.1 models (Nov 2025) - LATEST & BEST
+    GrokModel["GROK_4_1"] = "grok-4.1";
+    GrokModel["GROK_4_1_FAST"] = "grok-4.1-fast";
+    // Previous fast models (2025) - Still good
     GrokModel["CODE_FAST"] = "grok-code-fast-1";
     GrokModel["GROK_4_FAST_REASONING"] = "grok-4-fast-reasoning";
     GrokModel["GROK_4_FAST"] = "grok-4-fast-non-reasoning";
@@ -28,7 +31,7 @@ export var GrokModel;
 /**
  * Call Grok API
  */
-export async function callGrok(messages, model = GrokModel.GROK_4_FAST_REASONING, // Changed: Use cheap fast reasoning by default
+export async function callGrok(messages, model = GrokModel.GROK_4_1, // Updated 2025-11-21: Use latest Grok 4.1 by default
 temperature = 0.7, maxTokens = 16384, // Increased default for comprehensive responses
 forceVisibleOutput = true) {
     if (!GROK_API_KEY) {
@@ -44,7 +47,9 @@ forceVisibleOutput = true) {
     });
     try {
         // For Grok 4 models, we need to handle reasoning tokens specially
-        const isGrok4 = model === GrokModel.GROK_4_FAST_REASONING ||
+        const isGrok4 = model === GrokModel.GROK_4_1 ||
+            model === GrokModel.GROK_4_1_FAST ||
+            model === GrokModel.GROK_4_FAST_REASONING ||
             model === GrokModel.GROK_4_FAST ||
             model === GrokModel.GROK_4_HEAVY;
         // Adjust prompt for Grok 4 to ensure visible output
@@ -113,7 +118,7 @@ export const grokReasonTool = {
         const messages = [
             {
                 role: "system",
-                content: `You are Grok, an expert at logical reasoning and problem-solving.
+                content: `You are Grok 4.1, an expert at logical reasoning and problem-solving with enhanced emotional intelligence.
 ${approachPrompts[approach]}.
 ${context ? `Context: ${context}` : ''}`
             },
@@ -122,8 +127,8 @@ ${context ? `Context: ${context}` : ''}`
                 content: problem
             }
         ];
-        // Use GROK_4_FAST_REASONING by default (3x cheaper!), GROK_4_HEAVY only if explicitly requested
-        const model = useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_FAST_REASONING;
+        // Use GROK_4_1 by default (latest with enhanced reasoning!), GROK_4_HEAVY only if explicitly requested
+        const model = useHeavy ? GrokModel.GROK_4_HEAVY : GrokModel.GROK_4_1;
         const maxTokens = useHeavy ? 100000 : 16384; // 100k for heavy, 16k for normal reasoning
         log?.info(`Using Grok model: ${model} for deep reasoning (max tokens: ${maxTokens}, cost: ${useHeavy ? 'expensive $3/$15' : 'cheap $0.20/$0.50'})`);
         return await callGrok(messages, model, 0.7, maxTokens, true);
@@ -154,7 +159,7 @@ export const grokCodeTool = {
         const messages = [
             {
                 role: "system",
-                content: `You are Grok Code Fast, an expert programmer and code analyst.
+                content: `You are Grok 4.1 Fast, an expert programmer and code analyst with advanced tool-calling capabilities.
 Task: ${taskPrompts[task]}
 ${language ? `Language: ${language}` : ''}
 ${requirements ? `Requirements: ${requirements}` : ''}`
@@ -164,8 +169,8 @@ ${requirements ? `Requirements: ${requirements}` : ''}`
                 content: `Code:\n\`\`\`${language || ''}\n${code}\n\`\`\``
             }
         ];
-        log?.info(`Using Grok Code Fast (92 tok/sec coding specialist)`);
-        return await callGrok(messages, GrokModel.CODE_FAST, 0.2, 4000, true);
+        log?.info(`Using Grok 4.1 Fast (2M context, enhanced reasoning, $0.20/$0.50)`);
+        return await callGrok(messages, GrokModel.GROK_4_1_FAST, 0.2, 4000, true);
     }
 };
 /**

package/dist/src/workflows/model-router.js

@@ -43,6 +43,22 @@ export class ModelRouter {
                     reasoning: 10,
                     useFor: ['complex reasoning with evidence']
                 }],
+            ['grok-4.1', {
+                    id: 'grok-4.1',
+                    cost: 9,
+                    quality: 10,
+                    speed: 7,
+                    reasoning: 10,
+                    useFor: ['enhanced reasoning', 'creativity', 'emotional intelligence', 'first-principles']
+                }],
+            ['grok-4.1-fast', {
+                    id: 'grok-4.1-fast',
+                    cost: 9,
+                    quality: 10,
+                    speed: 8,
+                    reasoning: 10,
+                    useFor: ['tool-calling', 'agentic workflows', 'code analysis', 'fast reasoning']
+                }],
             ['grok-4', {
                     id: 'grok-4',
                     cost: 9,
@@ -134,14 +150,14 @@ export class ModelRouter {
             return 'lmstudio-local';
         }
         const taskTypeMap = {
-            'code': task.complexity > 0.7 ? 'qwen3-coder-480b' : 'gemini-2.5-flash',
+            'code': task.complexity > 0.7 ? 'qwen3-coder-480b' : 'grok-4.1-fast',
             'research': 'perplexity-sonar-pro',
             'reasoning': task.complexity > 0.5 ? 'gpt5' : 'gpt5_mini',
             'scout': 'multi-model',
             'verifier': task.complexity > 0.5 ? 'gpt5' : 'gpt5_mini',
             'challenger': 'gpt5_mini',
             'auditor': 'perplexity-sonar-pro',
-            'architect': 'grok-4',
+            'architect': 'grok-4.1',
             'commit_guardian': 'gemini-2.5-flash'
         };
         return taskTypeMap[task.type] || this.selectByConstraints(task, constraints);

package/docs/API_KEYS.md

@@ -116,7 +116,9 @@ Grok (by xAI) provides live web search, reasoning, and code analysis.
 
 #### Models Available
 
-- **grok-4** - Latest reasoning model
+- **grok-4.1** - Latest (Nov 2025): Enhanced reasoning, creativity & emotional intelligence
+- **grok-4.1-fast** - Tool-calling optimized: Fast inference, agentic workflows
+- **grok-4** - Previous reasoning model
 - **grok-4-0709** - Specific version
 - **grok-4-heavy** - Extended context (256k tokens)
 
@@ -124,7 +126,9 @@ Grok (by xAI) provides live web search, reasoning, and code analysis.
 
 | Model | Input | Output | Notes |
 |-------|-------|--------|-------|
-| Grok-4 | $5.00 / 1M tokens | $15.00 / 1M tokens | Standard |
+| Grok-4.1 | $0.20 / 1M tokens | $0.50 / 1M tokens | Latest & best! |
+| Grok-4.1-fast | $0.20 / 1M tokens | $0.50 / 1M tokens | Tool-calling optimized |
+| Grok-4 | $5.00 / 1M tokens | $15.00 / 1M tokens | Previous version |
 | Grok-4-heavy | $10.00 / 1M tokens | $30.00 / 1M tokens | 256k context |
 | **Live Search** | **$5 / 1k sources** | - | Extra cost per search! |
 

package/docs/TOOLS_REFERENCE.md

@@ -514,7 +514,7 @@ perplexity_reason({
 
 ### grok_search
 
-Cost-optimized web search using Grok-4's live search with advanced filtering.
+Cost-optimized web search using Grok-4.1's live search with advanced filtering and enhanced reasoning.
 
 #### Schema
 
@@ -640,7 +640,7 @@ grok_search({
 
 ### grok_reason
 
-Deep logical reasoning with Grok-4 using first principles.
+Deep logical reasoning with Grok-4.1 using first principles and enhanced emotional intelligence.
 
 #### Schema
 
@@ -686,7 +686,7 @@ grok_reason({
 
 ### grok_code
 
-Code analysis and optimization with Grok-4.
+Code analysis and optimization with Grok-4.1 Fast (tool-calling optimized).
 
 #### Schema
 
@@ -738,7 +738,7 @@ grok_code({
 
 ### grok_debug
 
-Deep debugging assistance with Grok-4.
+Deep debugging assistance with Grok-4.1 Fast.
 
 #### Schema
 
@@ -772,7 +772,7 @@ grok_debug({
 
 ### grok_architect
 
-System architecture and design with Grok-4.
+System architecture and design with Grok-4.1.
 
 #### Schema
 
@@ -800,7 +800,7 @@ grok_architect({
 
 ### grok_brainstorm
 
-Creative brainstorming using Grok-4-heavy.
+Creative brainstorming using Grok-4.1 with enhanced creativity and emotional intelligence.
 
 #### Schema
 

package/package.json

@@ -2,7 +2,7 @@
   "name": "tachibot-mcp",
   "mcpName": "io.github.byPawel/tachibot-mcp",
   "displayName": "TachiBot MCP - Universal AI Orchestrator",
-  "version": "2.0.5",
+  "version": "2.0.6",
   "type": "module",
   "main": "dist/src/server.js",
   "bin": {

package/smithery.yaml

@@ -1,7 +1,7 @@
 # Smithery configuration file: https://smithery.ai/docs/config#smitheryyaml
 name: "tachibot-mcp"
-description: "Multi-model AI orchestration platform with 26 tools across Perplexity, Grok, OpenAI GPT-5, Gemini, and Qwen. Features YAML workflows, 5 token-optimized profiles (4k-19k), smart routing, and cost controls. Build complex AI pipelines with variable passing and parallel execution."
-version: "2.0.0"
+description: "Multi-model AI orchestration platform with 31 tools across Perplexity, Grok, OpenAI GPT-5, Gemini, and Qwen. Features YAML workflows, 5 token-optimized profiles (4k-20k), smart routing, and cost controls. Build complex AI pipelines with variable passing and parallel execution."
+version: "2.0.6"
 
 startCommand:
   type: stdio