MCP4MCP Server

mcp4mcp is an AI-powered development assistant that tracks your MCP tools, detects duplicates, and provides guided suggestions throughout your MCP project. It helps you stay organized, reuse existing work, and accelerate your progress by offering intelligent insights as you build MCP servers and their tools.

How to use

You use mcp4mcp by connecting it to an MCP client or an integrated development workflow. Start a development session to log what you are building, then let it scan your codebase to discover existing tools and check for duplicates before you add new features. Ask for AI guidance at key points to get actionable next steps, and continuously track progress across your development sessions.

How to install

Prerequisites: Python is required on your system. You also need a Python package execution tool for zero-install usage if you prefer instant execution.

# Prerequisites
- Ensure Python is installed: https://www.python.org/downloads/
- Install pip if it is not already available

Zero-Install Setup is recommended for quick starts. Use uvx to run mcp4mcp without a local installation.

{
  "mcpServers": {
    "mcp4mcp": {
      "command": "uvx",
      "args": ["--from", "mcp4mcp", "mcp4mcp-server"]
    }
  }
}

Configuration and usage notes

You can run mcp4mcp in multiple ways. For development you may run a local Python server or use a zero-install approach with uvx. The following are the patterns described for connecting and using mcp4mcp with an MCP client or Claude Desktop.

Example local development server invocation with Python (manual usage): start a local server via Python when testing locally.

# Example local server startup for testing
python server.py

Zero-Install setup with uvx lets you run the mcp4mcp-server without a prior installation. This is convenient for quick experiments and demos.

Examples and workflows

Use mcp4mcp alongside Claude Desktop to enable AI-powered MCP development. Start the server, then configure Claude Desktop to connect to it, and begin interacting to build MCP tools with built-in guidance.

Example workflow highlights include automatic session tracking, duplicate detection, AI-driven next actions, and progress analytics across your development sessions.

Troubleshooting

If you encounter issues, ensure dependencies are installed and that the server is reachable. Validate Python imports and test the server startup command in your environment.

# Quick checks
python -c "import mcp4mcp; print('OK')"
# Start server for testing
python server.py

Development and testing

For local development, clone the project, install dependencies, and run tests. This workflow lets you iterate quickly while keeping the MCP tools in sync.

# Development setup
git clone https://github.com/hmatt1/mcp4mcp.git
cd mcp4mcp
pip install -e .[dev]

# Run tests
pytest -q

Using with Claude Desktop

Claude Desktop can automatically leverage mcp4mcp to track development sessions, detect duplicates, and surface AI guidance. Configure a MCP server entry pointing to the mcp4mcp server, then chat with Claude Desktop about your MCP project.

Core concepts and tools

The MCP server exposes tools to load project state, scan files for tools, check for conflicts before building, suggest next actions, track sessions, and provide analytics.

AI-powered development prompts

Use structured prompts to guide your development. Start sessions, check for duplicates, request AI suggestions, update progress, discover existing tools, and review analytics to improve your workflow.

Available tools

track_development_session_tool

Log your development sessions to monitor progress and usage over time.

update_project_state_tool

Add or update tools and project state as you build, marking stages like planned, in_progress, and completed.

scan_project_files_tool

Automatically discover tools in your codebase to promote reuse and reduce duplication.

check_before_build_tool

Check for conflicts or duplicates before starting to build a new tool.

suggest_next_action_tool

Provide AI-powered recommendations on what to build next based on current state.

analyze_tool_similarity_tool

Find similar tools to consolidate functionality and avoid duplication.

get_session_analytics_tool

View analytics about your development sessions and tool usage.