Docs MCP Server

🚀 MCP Document Search Server

The MCP Document Search Server is a lightweight MCP server designed to search and retrieve relevant document content from popular AI libraries such as LangChain, LlamaIndex, and OpenAI. It serves as a bridge between LLMs and external document sources, enabling language models to dynamically query and access the latest document content.

🚀 Quick Start

Setup Steps

1. Clone the repository:

   git clone https://github.com/your-repository.git
   

2. Create and activate a virtual environment:

   python -m venv env
   source env/bin/activate  # Use `env\Scripts\activate` on Windows
   

3. Install dependencies:

   pip install serper BeautifulSoup4
   

4. Configure environment variables:
   • Create a .env file and add:

     SERPER_API_KEY=your_api_key_here
     

Claude Desktop Integration

1. Modify the configuration file config.json:

   {
     "inspectorEnabled": true,
     "InspectableProcesses": [
       {
         "ProcessName": "uv",
         "PortNumber": 0
       }
     ]
   }
   

2. Add an example code block:

   {
     \"tool\": {
       \"name\": \"get_docs\",
       \"description\": \"Search for relevant content from the specified document library\",
       \"parameters\": {
         \"type\": \"object\",
         \"properties\": {
           \"query\": {
             \"type\": \"string\",
             \"description\": \"Search query string\"
           },
           \"filter\": {
             \"type\": \"string\",
             \"description\": \"Document filtering criteria\"
           }
         }
       }
     }
   }
   

Debugging

Use Node.js and the Inspector tool for real-time debugging:

npx @modelcontextprotocol/inspector uv run main.py

✨ Features

1. Web Search Integration

• Integrate web search using the Serper API.
• Support retrieving relevant information from multiple sources.

2. Content Extraction

• Parse HTML documents using BeautifulSoup.
• Provide clean and structured text output.

3. Tool Interface

• get_docs tool: Parameters include a query string and document filtering criteria.
• Workflow:
  • Search for relevant documents.
  • Parse the retrieved content.
  • Return the results.

📚 Documentation

MCP Protocol

The MCP protocol is an open standard for building secure two-way connections that link data sources with AI tools. Here are the core elements of the protocol:

Graph Structure

• LLM in isolation: The language model operates without tool support.
• LLM with tool support: The language model can call tools for information retrieval or processing.
• MCP environment: The language model interacts with external tools and data sources through the MCP protocol.

Ecosystem Roles

• Agent: Responsible for coordinating communication between the LLM and external tools.
• Document source: Provides searchable content, such as API documentation and knowledge bases.

Purpose and Vision

The goal of MCP is to standardize communication between language models and external tools and reduce the effort required for customized integration. Its vision is to promote the development of a modular service ecosystem, similar to a plugin architecture.

🔧 Technical Details

Supported Libraries/Documents

• LangChain
• LlamaIndex
• OpenAI

Future Improvement Directions

• Expand support for more libraries (e.g., HuggingFace, PyTorch, TensorFlow).
• Implement content caching to improve performance.
• Introduce a scoring mechanism to optimize the quality of search results.
• Enhance testing and exception handling capabilities to ensure stability in a production environment.