Vibe MCP Server

A Model Context Protocol (MCP) server that provides AI-integrated access to vibe-cli for unified AI coding rules management.

Features

📊 Deep Project Analysis - Comprehensive tech stack detection and AI configuration analysis
🚀 Rules Management - Initialize, sync, and manage AI coding standards across platforms
📈 Performance Monitoring - Real-time status and health checks for all AI platforms
🤖 AI-Optimized Output - Structured JSON responses designed for AI consumption

Demo

Repository Analysis with Vibe

See how AI assistants analyze projects and manage unified coding standards

Installation

One-Click Install

VS Code

VS Code Insiders

Claude Desktop Configuration

Add to your Claude Desktop config:

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

{
  "mcpServers": {
    "vibe": {
      "command": "npx",
      "args": ["-y", "@jinjos/vibe-mcp"],
      "cwd": "/path/to/your/project"
    }
  }
}

Cursor/VS Code with MCP

{
  "mcpServers": {
    "vibe": {
      "command": "npx", 
      "args": ["-y", "@jinjos/vibe-mcp"]
    }
  }
}

Available Tools

`repo_analysis`

Performs deep analysis of project tech stack, AI configurations, and rule coverage.

{
  "projectPath": "/path/to/project",
  "options": {
    "verbose": true
  }
}

Returns AI-optimized JSON with:

Project metadata and tech stack breakdown
AI platform configuration status
Coverage score with detailed analysis
Gap analysis with actionable recommendations
MCP server recommendations based on actual tech stack

`init_rules`

Sets up unified AI rules system with platform configurations.

{
  "projectPath": "/path/to/project",
  "options": {
    "platforms": "cursor,claude,copilot,gemini",
    "full": false
  }
}

`sync_rules`

Syncs rules from platform directories to unified vibe/ system.

{
  "projectPath": "/path/to/project",
  "options": {
    "platforms": "auto",
    "full": false,
    "dryRun": true
  }
}

`vibe_status`

Comprehensive status of vibe system including performance metrics.

{
  "projectPath": "/path/to/project",
  "options": {
    "verbose": true
  }
}

Requirements

Node.js 18.0.0 or higher
An MCP-compatible client (Claude Desktop, VS Code, Cursor, etc.)

Architecture

AI Agent ↔ MCP Protocol ↔ Vibe MCP Server ↔ vibe-cli

The MCP server provides:

Clean JSON output without emojis or formatting
Structured data optimized for AI consumption
Direct integration with vibe-cli functionality
Performance metrics and health monitoring

Configuration

Environment Variables

VIBE_LOG_LEVEL: Logging level (0=ERROR, 1=WARN, 2=INFO, 3=DEBUG)

Debugging

Enable debug logging:

VIBE_LOG_LEVEL=3 npx -y @jinjos/vibe-mcp

Troubleshooting

Tool not found errors

npm run build  # Ensure server is built

MCP connection issues

# Enable debug logging
VIBE_LOG_LEVEL=3 node dist/index.js

Performance monitoring

The server maintains:

Memory usage <100MB baseline
Response time <2 seconds for most operations
No background processes

Related Projects

vibe-cli - Universal AI rules CLI
Model Context Protocol - MCP specification
MCP Servers - Official MCP server registry

License

MIT - Same as vibe-cli

🤖 Bring your AI tools into perfect harmony 🤖