Skip to main content
Glama

MCP VectorStore Server

by BierschneiderEmanuel
GitHub
Python
MIT License
  • Linux
  • Apple
OverviewInspectNewEndpointsSchemaRelated ServersReviewsScore
Need Help?View Source CodeReport Issue

MCP VectorStore Server

A Model Context Protocol (MCP) server that provides advanced vector store operations for document search, PDF processing, and information retrieval. This server wraps the functionality from vectorstore.py into a standardized MCP interface.

Features

  • Vector Store Operations: Create, search, and manage document vector stores

  • PDF Processing: Extract and index content from PDF documents using LLMSherpa

  • Semantic Search: Advanced document search using HuggingFace embeddings

  • Web Search Integration: Google, Wikipedia, and DuckDuckGo search capabilities

  • File Operations: Read and process local files

  • Mathematical Calculations: Built-in calculator functionality

Prerequisites

System Requirements

  • Python: 3.8 or higher

  • Operating System: Linux, macOS, or Windows

  • Memory: Minimum 4GB RAM (8GB+ recommended for large document collections)

  • Storage: At least 2GB free space for models and vector stores

  • Network: Internet connection for downloading models and web searches

Optional GPU Support

For improved performance with large document collections:

  • CUDA: 11.8 or higher

  • GPU: NVIDIA GPU with 4GB+ VRAM

  • cuDNN: Compatible version for your CUDA installation

Installation

Step 1: Clone or Download the Repository

# If you have the files locally, navigate to the directory cd /path/to/McpDocServer # Or clone from a repository (if available) # git clone <repository-url> # cd McpDocServer

Step 2: Create a Virtual Environment

# Create a virtual environment python3 -m venv venv # Activate the virtual environment # On Linux/macOS: source venv/bin/activate # On Windows: # venv\Scripts\activate

Step 3: Install Dependencies

# Upgrade pip pip install --upgrade pip # Install all required packages pip install -r requirements.txt

Step 4: Install LLMSherpa (Optional but Recommended)

For optimal PDF processing, install LLMSherpa locally:

# Install LLMSherpa pip install llmsherpa # Start the LLMSherpa server (in a separate terminal) llmsherpa --port 5001

Step 5: Download Embedding Models

The server will automatically download the required embedding model on first use, but you can pre-download it:

# Download the embedding model python -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('sentence-transformers/all-mpnet-base-v2')"

Configuration

Environment Variables

Create a .env file in the project directory:

# LLMSherpa API URL (use local if available, otherwise cloud) LLMSHERPA_API_URL=http://localhost:5001/api/parseDocument?renderFormat=all # Vector store directory VECTORSTORE_DIR=/path/to/your/documents # User agent for web scraping USER_AGENT=Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/119.0.0.0 Safari/537.36 # Optional: CUDA device for GPU acceleration CUDA_VISIBLE_DEVICES=0

Directory Structure

Prepare your document directory:

your_documents/ ├── pdfs/ │ ├── document1.pdf │ ├── document2.pdf │ └── ... ├── text_files/ │ ├── notes.txt │ └── ... └── other_documents/ └── ...

Usage

Starting the MCP Server

# Make the server executable chmod +x mcp_vectorstore_server.py # Start the server on linux python /home/em/McpDocServer/mcp_vectorstore_server.py or windows with wsl wsl -d Ubuntu-24.04 bash -c "/mnt/c/Users/emanu/Desktop/McpDocServer/start_mcp.sh"

Using with MCP Clients

0. Claude Desktop

Add to your MCP configuration:

{ "mcpServers": { "vectorstore": { "command": "python", "args": ["/home/em/McpDocServer/mcp_vectorstore_server.py"], "env": { "PYTHONPATH": "/home/em/McpDocServer/McpDocServer" } } } }

1. GitHub Copilot

  1. Click on Configure Tools in the GitHub Copilot Chat window:

  2. Click on Add More Tools in the top search bar.

  3. Click on Add MCP Server in the top search bar.

  4. Click on command (stdio) in the top search bar.

  5. Enter command to run:

  6. python /home/em/McpDocServer/mcp_vectorstore_server.py or on windows: wsl -d Ubuntu-24.04 /mnt/c/Users/emanu/Desktop/McpDocServer/start_mcp.sh

  7. Enter mcp server id / name e.g. McpDocServer-19be5552

  8. Configure settings.json

{ "security.workspace.trust.untrustedFiles": "open", "python.defaultInterpreterPath": "/mnt/c/Users/emanu/Desktop/LLM/venv/venv/bin/python", "terminal.integrated.inheritEnv": false, "git.openRepositoryInParentFolders": "never", "terminal.integrated.scrollback": 100000, "mcp": { "servers": { "McpDocServer-19be5552": { "type": "stdio", "command": "python", "args": [ "/mnt/c/Users/emanu/Desktop/McpDocServer/mcp_vectorstore_server.py" ] } } } }

  1. Check if the following tools are available in the mcp server tool list when you click on Configure Tools in the GitHub Copilot Chat window and scroll to bottom:     vectorstore_search     vectorstore_create     vectorstore_info     vectorstore_clear     read_file     google_search     wikipedia_search     duckduckgo_search     calculate

  2. Select Agent mode in GitHub Copilot Chat window and use vectorstore_search to get information: use vectorstore_search to get information on unit testing 11)Confirm tool call usage.

2. Continue MCP CLient

name: McpDocServer version: 1.0.1 schema: v1 mcpServers: - name: McpDocServer command: wsl -d Ubuntu-24.04 args: - "/mnt/c/Users/emanu/Desktop/McpDocServer/start_mcp.sh" env: {} mcp_timeout: 180 # set timeout to 180 sec timeout: 9999 connectionTimeout: 120000 # 120 seconds = 2 minutes

3. Other MCP Clients

Configure your MCP client to use the server:

# Example with a generic MCP client mcp-client --server python --args /path/to/McpDocServer/mcp_vectorstore_server.py

Available Tools

Vector Store Operations

vectorstore_search

Search the vector store for relevant documents.

Parameters:

  • query (string, required): Search query

  • k (integer, optional): Number of results (default: 2)

Example:

{ "name": "vectorstore_search", "arguments": { "query": "machine learning algorithms", "k": 5 } }

vectorstore_create

Create a new vector store from documents in a directory.

Parameters:

  • directory_path (string, required): Path to directory containing documents

Example:

{ "name": "vectorstore_create", "arguments": { "directory_path": "/home/user/documents/research_papers" } }

vectorstore_info

Get information about the current vector store.

Example:

{ "name": "vectorstore_info", "arguments": {} }

vectorstore_clear

Clear all documents from the vector store.

Example:

{ "name": "vectorstore_clear", "arguments": {} }

File Operations

read_file

Read the contents of a file on the system.

Parameters:

  • filename (string, required): Path to the file to read

Example:

{ "name": "read_file", "arguments": { "filename": "/home/user/documents/notes.txt" } }

Web Search Operations

google_search

Search Google for information.

Parameters:

  • query (string, required): Search query

  • max_results (integer, optional): Maximum number of results (default: 3)

Example:

{ "name": "google_search", "arguments": { "query": "latest AI developments 2024", "max_results": 5 } }

wikipedia_search

Search Wikipedia for information.

Parameters:

  • query (string, required): Search query

Example:

{ "name": "wikipedia_search", "arguments": { "query": "artificial intelligence" } }

duckduckgo_search

Search DuckDuckGo for information.

Parameters:

  • query (string, required): Search query

Example:

{ "name": "duckduckgo_search", "arguments": { "query": "privacy-focused search engines" } }

Utility Operations

calculate

Perform mathematical calculations.

Parameters:

  • operation (string, required): Mathematical operation to perform

Example:

{ "name": "calculate", "arguments": { "operation": "2 + 2 * 3" } }

Resources

The server provides the following resources:

vectorstore://info

Returns information about the current vector store in JSON format.

Example Response:

{ "num_documents": 150, "directory": "/home/user/documents", "embeddings_model": "sentence-transformers/all-mpnet-base-v2" }

Troubleshooting

Common Issues

1. Import Errors

Problem: ModuleNotFoundError for various packages Solution: Ensure all dependencies are installed:

pip install -r requirements.txt

2. CUDA/GPU Issues

Problem: CUDA-related errors Solution: Install CPU-only versions:

pip uninstall faiss-gpu torch pip install faiss-cpu

3. LLMSherpa Connection Issues

Problem: Cannot connect to LLMSherpa API Solution:

  • Start LLMSherpa server: llmsherpa --port 5001

  • Or use cloud API by updating the URL in the code

4. Memory Issues

Problem: Out of memory errors with large documents Solution:

  • Reduce chunk size in the text splitter

  • Use smaller embedding models

  • Process documents in batches

5. Permission Issues

Problem: Cannot read files or directories Solution: Check file permissions:

chmod 644 /path/to/documents/* chmod 755 /path/to/documents/

Performance Optimization

For Large Document Collections

  1. Use GPU acceleration:

    # In vectorstore.py, ensure CUDA is enabled model_kwargs={'device': 'cuda'}

  2. Optimize chunk size:

    # Adjust in PDFVectorStoreTool.__init__ chunk_size=1000, # Smaller chunks for better performance chunk_overlap=100,

  3. Batch processing:

    # Process documents in smaller batches batch_size = 10

For Better Search Results

  1. Adjust similarity threshold:

    # In vectorstore_search method similarity_threshold = 0.7

  2. Use different embedding models:

    # Try different models for better results model_name="sentence-transformers/all-MiniLM-L6-v2" # Faster model_name="sentence-transformers/all-mpnet-base-v2" # Better quality

Development

Project Structure

McpDocServer/ ├── mcp_vectorstore_server.py # Main MCP server ├── vectorstore.py # Original vectorstore implementation ├── requirements.txt # Python dependencies ├── README.md # This documentation └── .env # Environment variables (create this)

Contributing

  1. Fork the repository

  2. Create a feature branch

  3. Make your changes

  4. Add tests if applicable

  5. Submit a pull request

Testing

# Run basic functionality tests python -c " from mcp_vectorstore_server import * print('Server imports successfully') " # Test vector store operations python -c " from vectorstore import PDFVectorStoreTool tool = PDFVectorStoreTool() print(f'Vector store initialized with {tool.vectorstore_get_num_items()} documents') "

License

This project is provided as-is for educational and research purposes. Please ensure you comply with the licenses of all included dependencies.

Support

For issues and questions:

  1. Check the troubleshooting section above

  2. Review the error logs

  3. Ensure all dependencies are correctly installed

  4. Verify your system meets the requirements

Changelog

Version 1.0.0

  • Initial release

  • MCP server implementation

  • Vector store operations

  • Web search integration

  • File operations

  • Mathematical calculations

This server cannot be installed

-
security - not tested
A
license - permissive license
-
quality - not tested

How are these scores calculated?

hybrid server

The server is able to function both locally and remotely, depending on the configuration or use case.

Provides advanced document search and processing capabilities through vector stores, including PDF processing, semantic search, web search integration, and file operations. Enables users to create searchable document collections and retrieve relevant information using natural language queries.

  1. Features
    1. Prerequisites
      1. System Requirements
      2. Optional GPU Support
    2. Installation
      1. Step 1: Clone or Download the Repository
      2. Step 2: Create a Virtual Environment
      3. Step 3: Install Dependencies
      4. Step 4: Install LLMSherpa (Optional but Recommended)
      5. Step 5: Download Embedding Models
    3. Configuration
      1. Environment Variables
      2. Directory Structure
    4. Usage
      1. Starting the MCP Server
      2. Using with MCP Clients
    5. Available Tools
      1. Vector Store Operations
      2. File Operations
      3. Web Search Operations
      4. Utility Operations
    6. Resources
      1. vectorstore://info
    7. Troubleshooting
      1. Common Issues
      2. Performance Optimization
    8. Development
      1. Project Structure
      2. Contributing
      3. Testing
    9. License
      1. Support
        1. Changelog
          1. Version 1.0.0

        Related MCP Servers

        • Simple Document Processing MCP Server

          cablate
          A
          security
          A
          license
          A
          quality
          Provides comprehensive document processing, including reading, converting, and manipulating various document formats with advanced text and HTML processing capabilities.
          Last updated -
          16
          24
          15
          MIT License

        • RAG Documentation MCP Server

          jumasheff
          -
          security
          A
          license
          -
          quality
          Provides tools for retrieving and processing documentation through vector search, enabling AI assistants to augment their responses with relevant documentation context.
          Last updated -
          22
          MIT License
          • Apple

        • FastMCP Document Analyzer

          Tathagat017
          -
          security
          F
          license
          -
          quality
          A comprehensive document analysis server that performs sentiment analysis, keyword extraction, readability scoring, and text statistics while providing document management capabilities including storage, search, and organization.
          Last updated -
          • Apple
          • Linux

        • AVS Document Search System

          patw
          -
          security
          A
          license
          -
          quality
          A vector search system that enables semantic retrieval of document chunks using MongoDB Atlas Vector Search and Voyage AI embeddings, allowing users to search documents by meaning rather than just keywords.
          Last updated -
          2
          MIT License

        View all related MCP servers

        New MCP Servers

        MCP directory API

        We provide all the information about MCP servers via our MCP API.

        curl -X GET 'https://glama.ai/api/mcp/v1/servers/BierschneiderEmanuel/McpDocServer'

        If you have feedback or need assistance with the MCP directory API, please join our Discord server