mcp-server-vectorsearch

Vector Search MCP Server

Node.js server implementing Model Context Protocol (MCP) for local vector search operations.

Features

• PDF and text document processing
• Document chunking and vectorization
• Local embedding generation using transformers.js
• Vector similarity search with FAISS
• Operates entirely on the local machine

Note: The server will only allow operations on files and documents specified via args.

API

Resources

• vector://search: Vector search operations interface

Tools

• index_file

  • Index a file for vector search
  • Input: path (string)
  • Handles PDF and text files

• index_directory

  • Index all files in a directory recursively
  • Input:
    • path (string)
    • extensions (string[]): File extensions to index, defaults to ['.txt', '.pdf']
    • excludePatterns (string[]): Patterns to exclude

• search

  • Perform semantic search with a query
  • Input:
    • query (string): Search query
    • numResults (number): Number of results to return (default: 5)
    • threshold (number): Similarity threshold (default: 0.7)

• clear_index

  • Clear the current vector index
  • Input: None

• get_index_stats

  • Get statistics about the current index
  • Input: None
  • Returns:
    • Total documents indexed
    • File types
    • Index size

• list_indexed_files

  • List all files that have been indexed
  • Input: None

Usage with Claude Desktop

Add this to your claude_desktop_config.json:

NPX

{
  "mcpServers": {
    "vectorsearch": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-vectorsearch",
        "/path/to/documents/directory"
      ]
    }
  }
}

Docker

{
  "mcpServers": {
    "vectorsearch": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "--mount", "type=bind,src=/path/to/documents/directory,dst=/data/documents",
        "mcp/vectorsearch",
        "/data/documents"
      ]
    }
  }
}

Usage with VS Code

For quick installation, click the installation buttons below:

For manual installation, add the following JSON block to your User Settings (JSON) file in VS Code. You can do this by pressing Ctrl + Shift + P and typing Preferences: Open Settings (JSON).

Optionally, you can add it to a file called .vscode/mcp.json in your workspace. This will allow you to share the configuration with others.

NPX

{
  "mcp": {
    "servers": {
      "vectorsearch": {
        "command": "npx",
        "args": [
          "-y",
          "@modelcontextprotocol/server-vectorsearch",
          "${workspaceFolder}"
        ]
      }
    }
  }
}

Docker

{
  "mcp": {
    "servers": {
      "vectorsearch": {
        "command": "docker",
        "args": [
          "run",
          "-i",
          "--rm",
          "--mount", "type=bind,src=${workspaceFolder},dst=/data/documents",
          "mcp/vectorsearch",
          "/data/documents"
        ]
      }
    }
  }
}

Build

Docker build:

docker build -t mcp/vectorsearch -f Dockerfile .

License

This MCP server is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.