Overview
What is MCP-Agg?
MCP-Agg is a unified API service that provides consistent access to multiple tools and platforms, including GitHub and Slack, through a single interface. It is designed for developers and teams seeking workflow automation and enhanced productivity by integrating various services.
How to use MCP-Agg?
Install Python 3.12+, PostgreSQL, and the uv package manager. Clone the repository, create a virtual environment, install dependencies, copy .env.example to .env, and run alembic upgrade head. Start the service with uvicorn api.main:app --reload --port 8000 (development) or docker-compose up -d (Docker). Register and log in, connect your accounts for each supported platform, generate an MCP URL from the endpoint, and use it in your MCP client configuration.
Key features of MCP-Agg
- Unified tool interface across multiple platforms
- Secure authentication and authorization per service
- Extensible architecture to add new tools
- MCP client support with unique generated URLs
- Comprehensive documentation via Swagger UI and ReDoc
Use cases of MCP-Agg
- Manage GitHub repositories, issues, and pull requests from one API
- Post messages, reply to threads, and add reactions in Slack channels
- Automate cross-platform workflows by combining GitHub and Slack actions
- Access channel history and user profiles from Slack via a single endpoint
- Integrate multiple services into an MCP client for centralized tool access
FAQ from MCP-Agg
What platforms does MCP-Agg support?
It currently supports GitHub (list repos, manage issues/PRs, access user profiles) and Slack (list channels, post messages, reply to threads, add reactions, access channel history, retrieve user profiles).
What are the runtime requirements?
Python 3.12 or higher, a PostgreSQL database, and the uv package manager are required.
How do I run MCP-Agg in production?
Start the application with uvicorn api.main:app --host 0.0.0.0 --port 8000 or use docker-compose up -d for a containerized setup.
How can I test MCP-Agg?
Run tests using python -m pytest and generate coverage reports with python -m pytest --cov=api.
How do I connect an MCP client?
Register and log in, connect your platform accounts (GitHub, Slack, etc.), navigate to the MCP URL generator endpoint, and use the generated URL in your MCP client configuration.
