Simple File Vector Store

🚀 @lishenxydlgzs/simple-files-vectorstore

A Model Context Protocol (MCP) server that provides semantic search capabilities for files. This server monitors specified directories and generates vector embeddings of file contents, enabling semantic search for documents.

🚀 Quick Start

📦 Installation

Add the following to your MCP settings file:

{
  "mcpServers": {
    "files-vectorstore": {
      "command": "npx",
      "args": [
        "-y",
        "@lishenxydlgzs/simple-files-vectorstore"
      ],
      "env": {
        "WATCH_DIRECTORIES": "/path/to/your/directories"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

💻 Location of MCP Settings File

• VSCode Cline Extension: ~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json
• Claude Desktop Application: ~/Library/Application Support/Claude/claude_desktop_config.json

⚙️ Configuration

The server needs to be configured through environment variables.

📋 Required Environment Variables

You must specify the directories to monitor using one of the following methods:

• WATCH_DIRECTORIES: A comma-separated list of directories.
• WATCH_CONFIG_FILE: The path to a JSON configuration file containing a watchList array.

Example (Using WATCH_DIRECTORIES)

{
  "mcpServers": {
    "files-vectorstore": {
      "command": "npx",
      "args": [
        "-y",
        "@lishenxydlgzs/simple-files-vectorstore"
      ],
      "env": {
        "WATCH_DIRECTORIES": "/path/to/dir1,/path/to/dir2"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

Example (Using WATCH_CONFIG_FILE)

{
  "mcpServers": {
    "files-vectorstore": {
      "command": "npx",
      "args": [
        "-y",
        "@lishenxydlgzs/simple-files-vectorstore"
      ],
      "env": {
        "WATCH_CONFIG_FILE": "/path/to/watch-config.json"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

The watch configuration file should have the following structure:

{
  "watchList": [
    "/path/to/dir1",
    "/path/to/dir2",
    "/path/to/specific/file.txt"
  ]
}

🔢 Optional Environment Variables

• CHUNK_SIZE: The size of text chunks (default: 1000).
• CHUNK_OVERLAP: The amount of overlap between chunks (default: 200).
• IGNORE_FILE: The path to a file in .gitignore style for excluding files/directories.

Example (Including All Optional Parameters)

{
  "mcpServers": {
    "files-vectorstore": {
      "command": "npx",
      "args": [
        "-y",
        "@lishenxydlgzs/simple-files-vectorstore"
      ],
      "env": {
        "WATCH_DIRECTORIES": "/path/to/dir1,/path/to/dir2",
        "CHUNK_SIZE": "2000",
        "CHUNK_OVERLAP": "200",
        "IGNORE_FILE": "/path/to/ignore-file.txt"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

🛠️ MCP Tools

🔍 Search Tool

• Name: search
• Parameters:
  • query: The query string to search.
  • maxResults: The maximum number of returned results, default is 5.
• Response:

{
  "results": [
    {
      "id": string,
      "score": number,
      "content": string
    }
  ]
}

• Example:

curl -X POST http://localhost:1234/mcp/v1/tools/search \
  -H "Content-Type: application/json" \
  -d '{"query":"查询字符串", "maxResults":5}'

✨ Features

• Supports file content analysis for Large Language Models (LLMs).
• Provides efficient semantic search capabilities.
• Is scalable and supports indexing and retrieval of a large number of documents.