An intelligent MCP (Model Context Protocol) server that automatically generates and updates documentation based on code changes, designed to work directly with Cursor.

• AI-Powered Documentation: Uses Gemini to generate clear, accurate documentation
• Code Analysis: CodeRabbit structures and analyzes code changes
• Code Execution: Daytona executes code snippets in secure sandboxes
• Quality Evaluation: Galileo.ai evaluates documentation accuracy and tone
• Self-Correction: Automatic improvement loop when quality thresholds aren't met
• Git Integration: Automatic commits and version control
• MCP Integration: Works directly in Cursor via Model Context Protocol

pip install -r requirements.txt

Create a .env file with your API keys (optional - system works with fallbacks):

# Gemini Configuration (recommended)
GEMINI_API_KEY=your_gemini_api_key

# Daytona Configuration
DAYTONA_API_KEY=your_daytona_api_key
DAYTONA_API_URL=https://app.daytona.io/api

# Galileo Configuration
GALILEO_API_KEY=your_galileo_api_key
GALILEO_PROJECT_ID=your_project_id

# CodeRabbit Configuration
CODERABBIT_API_KEY=your_coderabbit_api_key

# Git Configuration
GIT_REPO_PATH=.
GIT_BRANCH=main

# Quality Thresholds
MIN_DOC_QUALITY_SCORE=0.7
MAX_SELF_CORRECTION_ATTEMPTS=3

Run the configuration helper:

python update_mcp_config.py

Or manually edit %USERPROFILE%\.cursor\mcp.json (Windows) or ~/.cursor/mcp.json (macOS/Linux):

{
  "mcpServers": {
    "live-document-editor": {
      "command": "python",
      "args": ["C:\\path\\to\\HackSprint\\mcp_server.py"],
      "env": {}
    }
  }
}

After updating the configuration, restart Cursor completely to load the MCP server.

Once connected, you can use these MCP tools directly in Cursor:

1. process_code_changes - Process Git code changes and generate documentation

   Use the process_code_changes tool to generate documentation for my recent code changes
   

2. get_git_diff - Get Git diff for the current repository

   Use the get_git_diff tool to show me the current Git diff
   

3. update_documentation - Update or create a documentation file

   Use the update_documentation tool to update README.md with [content]
   

4. commit_documentation - Commit documentation changes to Git

   Use the commit_documentation tool to commit the documentation
   

5. get_recent_commits - Get recent Git commits

   Use the get_recent_commits tool to show me the last 5 commits
   

Git Repository → MCP Server (mcp_server.py)
    ↓
    ├─ CodeRabbit: Structure code changes
    ├─ Gemini: Generate documentation
    ├─ Daytona: Execute code snippets
    ├─ Galileo: Evaluate documentation quality
    └─ Self-Correction Loop: Improve until quality threshold met
    ↓
Final Output → Git Repository (Commit to main)

.
├── mcp_server.py              # Main MCP server entry point (for Cursor)
├── src/
│   ├── orchestrator.py        # MCP orchestrator with Gemini
│   ├── git_handler.py         # Git integration
│   ├── config.py              # Configuration management
│   └── mcp_servers/
│       ├── coderabbit.py      # CodeRabbit server
│       ├── daytona_server.py  # Daytona server
│       └── galileo_server.py  # Galileo server
├── requirements.txt           # Dependencies
└── MCP_SETUP.md              # Detailed MCP setup guide

• Coordinates all MCP servers
• Uses Gemini for intelligent decision-making
• Manages self-correction loops

• Structures code changes from Git diffs
• Analyzes code complexity
• Categorizes modifications

• Executes code snippets in secure sandboxes
• Reports execution success/failure
• Handles code validation

• Evaluates documentation quality
• Assesses accuracy, tone, and clarity
• Provides feedback for improvement

• Monitors repository changes
• Commits documentation updates
• Manages version control

1. Code Changes Detected: Git diff triggers the system
2. Structure Analysis: CodeRabbit analyzes and structures the changes
3. Documentation Generation: Gemini generates initial documentation
4. Code Execution: Daytona executes code snippets to validate examples
5. Quality Evaluation: Galileo evaluates documentation quality
6. Self-Correction: If quality is low, the system self-corrects using Gemini
7. Final Output: High-quality documentation is committed to Git

The system includes fallback mechanisms:

• If Gemini is unavailable, basic documentation is generated
• CodeRabbit uses local parsing (no API required)
• Galileo uses local evaluation algorithms
• Daytona is optional for code execution

However, for best results, configure all API keys.

1. Check the path: Ensure the path in mcp.json is correct
2. Check Python path: Make sure python is in your PATH
3. Check dependencies: Ensure all dependencies are installed:

   pip install -r requirements.txt

4. Check logs: Look for errors in Cursor's Developer Console (Ctrl+Shift+I)

• Ensure the directory is a Git repository (git init)
• Check file permissions
• Verify Git is installed and in PATH

• MCP_SETUP.md - Detailed MCP setup and usage guide
• ARCHITECTURE.md - System architecture details
• SYSTEM_DESIGN.md - Workflow diagrams and design

MIT License