ckhamitkar/knowledge-base-mcp-server

Knowledge Base MCP Server

This is a knowledge base server that allows you to upload files, search them using semantic search, and ask questions about them using a LangChain agent.

Features

• File Management: Add, list, and read files in the knowledge base.
• Semantic Search: Search for documents using semantic similarity.
• LangChain Agent: Ask questions about the documents in the knowledge base.

Setup

1. Install dependencies:

   pip install -r requirements.txt
   

2. Set Hugging Face API Token: To use the LangChain agent, you need to set your Hugging Face API token as an environment variable:

   export HUGGINGFACEHUB_API_TOKEN="your-token-here"
   

3. Run the server:

   python app.py
   

   The server will be running at http://localhost:8000.

API

Files

• POST /files/: Upload a file.

  • Form Data: file
  • Example:

    curl -X POST -F "file=@/path/to/your/file.txt" http://localhost:8000/files/
    

• GET /files/: List all files.

  • Example:

    curl http://localhost:8000/files/
    

• GET /files/<filename>: Read a file.

  • Example:

    curl http://localhost:8000/files/your-file.txt
    

Search

• POST /index/: Re-index all documents.

  • Example:

    curl -X POST http://localhost:8000/index/
    

• POST /search/: Search for documents.

  • Body: {"query": "your search query"}
  • Example:

    curl -X POST -H "Content-Type: application/json" -d '{"query": "your search query"}' http://localhost:8000/search/
    

Agent

• POST /agent/: Ask a question to the agent.
  • Body: {"query": "your question"}
  • Example:

    curl -X POST -H "Content-Type: application/json" -d '{"query": "your question"}' http://localhost:8000/agent/