Sourcesync.ai

🚀 SourceSync.ai MCP Server Guide

SourceSync.ai MCP server is a powerful tool that provides document management and question - answering capabilities. It simplifies the process of handling documents and retrieving information from a knowledge base.

🚀 Quick Start

📦 Installation

To install the necessary dependencies, run the following command in the project root directory:

npm install

⚙️ Starting the Server

After building the project, start the server:

node dist/index.js

🔧 Environment Variable Configuration

Create a .env file and add the following content:

PORT=3001 # The port on which the service runs, default is 3001
DEBUG=false # Whether to enable debug logs, default is false

✨ Features

🛠️ Configuring MCP

Creating a New Configuration File

Create a new file, for example, custom-config.ts, in the src/config/ directory:

export const config = {
  host: "localhost", // The host address to which the service binds, default is all interfaces
  port: 3001,        // The port on which the service runs, default is 3001
  namespace: "",     // The default namespace, leave empty if not used
  logLevel: "info"   // Log level: debug, info, warn, error, default is info
};

Installation and Startup

npm install
node dist/index.js

📋 Tool Overview

Document Management Tools

• getDocuments: Retrieve all documents under a specified namespace.
• queryDocuments: Retrieve matching documents based on query conditions.
• addDocument: Add a new document to the knowledge base.
• updateDocument: Update the content of an existing document.
• deleteDocument: Delete a specified document.

Question - Answering Tools

• answerQuestion: Answer questions based on the knowledge base content.
• contextWindow: Set the context window size to control the information retrieval scope.
• retrievalMode: Switch between different retrieval modes (exact, fuzzy).

📄 Document Retrieval Workflow

1. Use getDocuments and set includeConfig.parsedTextFileUrl: true to get documents containing content URLs.
2. Extract the URLs from the response.
3. Use fetchUrlContent to pull the actual content:

{
  "name": "fetchUrlContent",
  "arguments": {
    "url": "https://example.com"
  }
}

⚠️ Troubleshooting

Connection Issues

1. Verify Paths: Ensure that all paths in the configuration are absolute paths, not relative paths.
2. Check Permissions: Ensure that the server file has execution permissions (chmod +x dist/index.js).
3. Enable Developer Mode: Enable developer mode in Claude Desktop and view the MCP log file.
4. Run the Server Directly:

   node /path/to/sourcesyncai-mcp/dist/index.js
   

5. Restart the AI Client: After making changes, fully restart Claude Desktop or Cursor.
6. Check Environment Variables: Ensure that all required environment variables are correctly set.

Debug Logs

To record detailed information, add the DEBUG environment variable:

DEBUG=true

🔧 Technical Details

Project Structure

• src/index.ts: The main entry file and server configuration.
• src/schemas.ts: The schema definitions of all tools.
• src/sourcesync.ts: The client for interacting with the SourceSync.ai API.
• src/sourcesync.types.ts: TypeScript type definitions.

Building and Testing

# Build the project
npm run build

# Run tests
npm test

📄 License

This project is licensed under the MIT license.

🔗 Links

• SourceSync.ai Documentation
• SourceSync.ai API Reference
• Model Context Protocol