mcp_server

This MCP server provides a multi-feature backend that supports file operations, database access, API integration, and vector database interactions. It is designed to work with large language model integration workflows (such as Qwen) by offering a structured interface for uploading and managing files, connecting to MongoDB, calling external APIs, and storing/searching vectors. The server is designed to be deployed with Docker for easy one-click setup, and it includes a Python client and examples to demonstrate how to interact with it from a Qwen-based client. You can start the server, then use the included API endpoints or the Python client to upload files, persist and query vectors, and perform database operations.

Prerequisites:

• Docker and Docker Compose (for one-click deployment) or Python 3.10+ and pip
• Git
• Optional: MongoDB instance or accessible MongoDB URI

Installation steps (Docker-based):

1. Clone the repository: git clone https://github.com/ningwenjie/mcp_server cd mcp_server

2. Start the service with Docker Compose: docker-compose -f docker/docker-compose.yml up -d

3. Verify that the service is running: curl http://localhost:8000/health

Installation steps (Direct Python execution):

1. Create a Python virtual environment and install dependencies: python -m venv venv source venv/bin/activate pip install -r requirements.txt

2. Run the server directly: python main.py

3. Ensure configurations (e.g., MONGODB_URI) are set via environment variables or a config file before running.

Notes:

• Docker deployment is recommended for production to simplify dependencies and networking.
• If MongoDB or a vector store is not yet available, configure placeholders and update them once the services are accessible.

Tips and common issues:

• Ensure MongoDB URI (MONGODB_URI) is reachable from the MCP server container or environment where you run Python.
• When using Docker, expose port 8000 (or update the config) so health checks and clients can reach the API.
• The server supports file operations (upload, list, download, delete), vector storage and similarity search, API integrations, and MongoDB document operations.
• If you encounter API call failures, verify external API access permissions and required API keys (set in API_KEY or environment variables).
• For large vector dimensions, ensure the FAISS or chosen vector store backend is properly configured and has sufficient memory.