MCP Aggregated Tool Service
- python
3
GitHub Stars
python
Language
6 months ago
First Indexed
2 months ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
You deploy and run an MCP server that aggregates multiple tools behind a single API. It provides a standardized, asynchronous interface to connect, register, and invoke independent tools, making it easy to integrate with frontends or large language model workflows.
How to use
You access the MCP server at the defined endpoint and send tool requests via POST with a JSON body that matches the registered tool’s input model. The server automatically registers tools found in the tools directory and exposes them under the unified /mcp path. To use a tool, choose its name, prepare input according to its defined parameters, and submit a request to the MCP endpoint. You’ll receive a standardized JSON response with either a success result or an error message.
How to install
# Prerequisites
Python 3.8+
pip
Docker / Docker Compose (optional)
# Local development run (Python)
pip install -r requirements.txt
python main.py
# Default MCP URL
# http://localhost:8000/mcp
# Optional Docker deployment (standalone docker-compose setup)
docker-compose up --build -d
# Access the MCP URL after containers are up
# http://localhost:8000/mcp
Configuration and deployment notes
The MCP service is designed to start quickly in a local development environment and to be deployed with Docker. You can run it locally with Python or deploy via Docker Compose to manage containers in a consistent environment.
Security and maintenance
Ensure your MCP endpoint is accessible only to trusted clients in production. Consider network controls and authentication for tool access if you plan to expose the MCP API publicly. Regularly update dependencies and monitor tool registrations to maintain a secure and stable integration layer.
Deployment and runtime endpoints
The MCP server provides a single entry point for tool invocation under the /mcp path. In a local setup, you can reach the API at http://localhost:8000/mcp after starting the server. When running in Docker, the same endpoint is exposed through the container's mapped port.
