mcp-compose

MCP Compose provides a unified platform to manage multiple MCP servers through a modern web UI and REST API. It includes automatic discovery of MCP servers, intelligent service composition across tools, protocol translation between STDIO and SSE, and real-time monitoring. You can start multiple MCP servers, browse and invoke tools across them, and observe live logs and metrics from a single control plane. The CLI exposes commands to serve via a web UI, discover available servers, and invoke tools, while the Python API allows programmatic orchestration and tool invocation from code.

Prerequisites:

• Python 3.10+ and pip
• Git

Install from PyPI:

pip install mcp-compose

Or install from source:

git clone https://github.com/datalayer/mcp-compose.git
cd mcp-compose
pip install -e .

Usage without Docker:

# Start the server with Web UI (requires Python environment)
mcp-compose serve --config examples/mcp_compose.toml

If you prefer Docker-based deployment:

# Clone and run via docker-compose (includes Prometheus & Grafana)
git clone https://github.com/datalayer/mcp-compose.git
cd mcp-compose
docker-compose up -d

Access the UI and API:

• Web UI: http://localhost:8000
• API: http://localhost:8000/api/v1

Note: For production, follow the Deployment Guide in the repository to configure tokens, TLS, and scalable deployment.

Tips and common setup notes:

• If you are using the CLI, ensure your environment PATH includes the mcp-compose script generated by the installation.
• Enable security features such as token authentication and CORS in your config for production use.
• Use the docker-compose deployment for quick, production-like testing with Prometheus and Grafana dashboards.
• The system supports protocol translation (STDIO ↔ SSE) to bridge legacy MCP servers with modern clients.
• Check the REST API and Web UI for real-time metrics, logs, and health checks to diagnose issues quickly.