Grok Code MCP Server

Expert coding assistant powered by xAI's grok-code-fast-1 model

Features • Installation • Usage • Tools • Examples • API

🚀 Features

• 256K Context Window - Handle massive codebases and multi-file projects
• 4x Faster - Optimized for speed compared to competing models
• 1/10th Cost - Economical pricing at $0.20/1M input tokens
• Native Reasoning - Access to model's reasoning traces in real-time
• Tool Calling - Native function calling support for complex workflows
• Production Ready - Battle-tested for real-world coding tasks

📦 Installation

Option 1: Install from npm (Recommended)

npm install -g grok-code-mcp-server

Option 2: Install from GitHub

npm install -g github:yourusername/grok-code-mcp-server

Option 3: Local Development

git clone https://github.com/yourusername/grok-code-mcp-server.git
cd grok-code-mcp-server
npm install
npm run build
npm link

🔧 Configuration

1. Get your API Key

Get your xAI API key from x.ai/api

2. Configure MCP

Add to your Claude Desktop or compatible MCP client configuration:

Claude Desktop (claude_desktop_config.json)

Windows: %APPDATA%\Claude\claude_desktop_config.json
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Linux: ~/.config/claude/claude_desktop_config.json

{
  "mcpServers": {
    "grok-code": {
      "command": "npx",
      "args": ["-y", "grok-code-mcp-server"],
      "env": {
        "XAI_API_KEY": "your-xai-api-key-here"
      }
    }
  }
}

Alternative: Using .mcp.json in your project

{
  "mcpServers": {
    "grok-code": {
      "command": "npx",
      "args": ["-y", "grok-code-mcp-server"],
      "env": {
        "XAI_API_KEY": "your-xai-api-key-here"
      },
      "timeout": 45000,
      "autoApprove": [
        "grok_generate_code",
        "grok_refactor_code",
        "grok_analyze_code"
      ]
    }
  }
}

3. Environment Variables

| Variable | Description | Default | |----------|-------------|---------| | XAI_API_KEY | Your xAI API key (required) | - | | XAI_API_URL | API endpoint URL | https://api.x.ai/v1/chat/completions | | XAI_TIMEOUT | Request timeout in ms | 30000 |

🛠️ Available Tools

grok_generate_code

Generate production-ready code with tests and documentation.

Parameters:

• task (required): Description of code to generate
• language: Programming language (typescript, python, go, etc.)
• framework: Framework to use (React, FastAPI, Next.js, etc.)
• context: Additional context or requirements
• includeTests: Generate unit tests (default: false)
• includeComments: Include helpful comments (default: true)
• style: Coding style preferences

Example:

{
  "tool": "grok_generate_code",
  "arguments": {
    "task": "Create a React hook for managing localStorage",
    "language": "typescript",
    "framework": "React",
    "includeTests": true
  }
}

grok_refactor_code

Refactor existing code for better quality and maintainability.

Parameters:

• code (required): Code to refactor
• goal (required): Refactoring goal
• language: Programming language
• preserveBehavior: Ensure behavior remains unchanged (default: true)
• targetPatterns: Design patterns to apply

grok_analyze_code

Analyze code for bugs, performance issues, and quality.

Parameters:

• code (required): Code to analyze
• aspects: Areas to analyze (performance, security, maintainability, bugs, style, complexity)
• language: Programming language

grok_review_pr

Review pull request changes with detailed feedback.

Parameters:

• diff (required): Git diff to review
• context: PR description
• focusAreas: Specific areas to focus on
• standards: Coding standards to check

grok_explain_code

Explain code functionality with adjustable complexity.

Parameters:

• code (required): Code to explain
• targetAudience: Complexity level (beginner, intermediate, expert)
• focusOn: Specific aspects to focus on

grok_debug_code

Debug code and identify issues with fixes.

Parameters:

• code (required): Code with bugs
• error: Error message
• expectedBehavior: What the code should do
• language: Programming language

grok_optimize_code

Optimize code for performance or readability.

Parameters:

• code (required): Code to optimize
• targetMetric: What to optimize (speed, memory, readability, all)
• constraints: Constraints to maintain

📚 Examples

Generate a TypeScript Function

// Using with Claude
"Can you use grok to generate a TypeScript function that validates email addresses with comprehensive tests?"

// Direct tool call
{
  "tool": "grok_generate_code",
  "arguments": {
    "task": "Create a function to validate email addresses",
    "language": "typescript",
    "includeTests": true,
    "includeComments": true
  }
}

Refactor Legacy Code

{
  "tool": "grok_refactor_code",
  "arguments": {
    "code": "// your legacy code here",
    "goal": "Convert to modern async/await pattern",
    "language": "javascript",
    "preserveBehavior": true
  }
}

Analyze Code Quality

{
  "tool": "grok_analyze_code",
  "arguments": {
    "code": "// code to analyze",
    "aspects": ["performance", "security", "maintainability"],
    "language": "python"
  }
}

🔌 Integration with Claude

When using with Claude Desktop or Claude.ai, the tools are available directly:

Human: Can you help me create a React component for a todo list using the grok code server?