scmdr sourcesyncai mcp

WHAT IT CONNECTS

What scmdr sourcesyncai mcp gives an AI Agent

• Manage namespaces for organizing knowledge
• Ingest content from various sources (text, URLs, websites, external services)
• Retrieve, update, and manage documents stored in your knowledge base
• Perform semantic and hybrid searches against your knowledge base
• Access document content directly from parsed text URLs
• Manage connections to external services
• Default configuration support for seamless AI integration
• Open Cursor Settings
• Go to `Features > MCP Servers`
• Click `+ Add New MCP Server`
• Enter the following:
• Name: `sourcesyncai-mcp` (or your preferred name)
• Type: `command`
• Command: `env SOURCESYNCAI_API_KEY=your-api-key npx -y sourcesyncai-mcp`
• Locate the Claude Desktop configuration file:
• **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
• **Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
• **Linux**: `~/.config/Claude/claude_desktop_config.json`
• Edit the configuration file to add the SourceSync.ai MCP server:
• Save the configuration file and restart Claude Desktop
• `SOURCESYNC_API_KEY`: Your SourceSync.ai API key (required)
• `SOURCESYNC_NAMESPACE_ID`: Default namespace ID to use for operations
• `SOURCESYNC_TENANT_ID`: Your tenant ID (optional)
• `validate_api_key`: Validate a SourceSync.ai API key
• `create_namespace`: Create a new namespace
• `list_namespaces`: List all namespaces
• `get_namespace`: Get details of a specific namespace
• `update_namespace`: Update a namespace
• `delete_namespace`: Delete a namespace
• `ingest_text`: Ingest text content
• `ingest_urls`: Ingest content from URLs
• `ingest_sitemap`: Ingest content from a sitemap
• `ingest_website`: Ingest content from a website
• `ingest_notion`: Ingest content from Notion
• `ingest_google_drive`: Ingest content from Google Drive
• `ingest_dropbox`: Ingest content from Dropbox
• `ingest_onedrive`: Ingest content from OneDrive
• `ingest_box`: Ingest content from Box
• `get_ingest_job_run_status`: Get the status of an ingestion job run
• `getDocuments`: Retrieve documents with optional filters
• `updateDocuments`: Update document metadata
• `deleteDocuments`: Delete documents
• `resyncDocuments`: Resync documents
• `fetchUrlContent`: Fetch text content from document URLs
• `semantic_search`: Perform semantic search
• `hybrid_search`: Perform hybrid search (semantic + keyword)
• `create_connection`: Create a new connection to an external service
• `list_connections`: List all connections
• `get_connection`: Get details of a specific connection
• `update_connection`: Update a connection
• `revoke_connection`: Revoke a connection
• "Search my SourceSync knowledge base for information about machine learning."
• "Ingest this article into my SourceSync knowledge base: [URL]"
• "Create a new namespace in SourceSync for my project documentation."
• "List all the documents in my SourceSync namespace."
• "Get the text content of document [document_id] from my SourceSync namespace."
• **Verify Paths**: Ensure all paths in your configuration are absolute paths, not relative.
• **Check Permissions**: Ensure the server file has execution permissions (`chmod +x dist/index.js`).
• **Enable Developer Mode**: In Claude Desktop, enable Developer Mode and check the MCP Log File.
• **Test the Server**: Run the server directly from the command line:
• **Restart AI Client**: After making changes, completely restart Claude Desktop or Cursor.
• **Check Environment Variables**: Ensure all required environment variables are correctly set.
• `src/index.ts`: Main entry point and server setup
• `src/schemas.ts`: Schema definitions for all tools
• `src/sourcesync.ts`: Client for interacting with SourceSync.ai API
• `src/sourcesync.types.ts`: TypeScript type definitions
• [SourceSync.ai Documentation](https://sourcesync.ai)
• [SourceSync.ai API Reference](https://sourcesync.ai/api-reference/authentication)
• [Model Context Protocol](https://modelcontextprotocol.io/introduction)
• First, use `getDocuments` with `includeConfig.parsedTextFileUrl: true` to get documents with their content URLs
• Extract the URL from the document response
• Use `fetchUrlContent` to retrieve the actual content: