Avs Docs MCP

🚀 MCP Document Search System

A vector search system for document retrieval using MongoDB Atlas Vector Search and Voyage AI embeddings.

🚀 Quick Start

This is a vector search system designed for document retrieval, leveraging MongoDB Atlas Vector Search and Voyage AI embeddings. The sample data provided is tailored for Atlas Vector Search.

✨ Features

• Ingests and chunks markdown documents with hierarchical headers.
• Generates embeddings using Voyage AI's contextual embeddings API.
• Stores documents and embeddings in MongoDB with parent - child relationships.
• Provides a FastMCP server for semantic document search.
• Supports configurable vector dimensions and chunking strategies.

Available MCP Tools

The document search server offers the following tools:

1. search_documents_vector(query: str, limit: int = 5)
   • The primary search method utilizing vector similarity.
   • Returns document chunks with metadata and similarity scores.
   • Ideal for semantic/meaning - based queries.
2. search_documents_lexicaly(query: str, limit: int = 1)
   • A fallback search method using lexical/text matching.
   • Returns full parent documents with search scores.
   • Useful when vector search fails to find good matches.
3. get_parent_document(parent_id: str)
   • Retrieves the complete parent document by ID.
   • Returns the original content and file path.
   • Use after a search to obtain the full context for a chunk.

📦 Installation

Prerequisites

• Python 3.10+
• A MongoDB Atlas cluster with vector search enabled
• A Voyage AI API key

Steps

1. Clone the repository:

git clone https://github.com/patw/avs-document-search.git
cd avs-document-search

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file based on sample.env with your credentials.

💻 Usage Examples

Basic Usage

1. Ingest documents in the docs/ directory:

python ingest_docs.py

2. Run the search server:

python avs-mcp.py

Running the search server mainly verifies that your MongoDB URI is correct. You will need to integrate this MCP server with an MCP client such as Claude Desktop. Here's a sample config:

{
  "mcpServers": {
    "Atlas Vector Search Docs": {
      "command": "uv",
      "args": [
        "run",
        "--with",
        "fastmcp, pymongo, requests",
        "fastmcp",
        "run",
        "<path to>/avs-docs-mcp/avs-mcp.py"
      ]
    }
  }
}

📚 Documentation

Configuration

Copy sample.env to .env and edit it to configure the following:

• MongoDB connection string
• Database and collection names
• Voyage AI API key
• Vector dimensions (default is 256)

Future Improvements

• Implement hybrid search combining vector and text search using $rankFusion (when MongoDB 8.1 is GA on Atlas).
• Support additional file formats (PDF, Word, etc.) with Docling.

🤝 Contributing

Pull requests are welcome! For major changes, please open an issue first.

👨‍💻 Author

Pat Wendorf
pat.wendorf@mongodb.com
GitHub: patw

📄 License

MIT