CData Sync

Desktop Usage: Stdio Transport (Claude Desktop)

Thestdio transportis designed for local desktop usage with the Claude Desktop app. This is the recommended approach for individual developers.

{ "mcpServers": { "cdata-sync-server": { "command": "node", "args": ](https://www.cdata.com/sync/)["/absolute/path/to/cdata-sync-mcp-server/dist/index.js"], "env": { "MCP_TRANSPORT_MODE": "stdio", "CDATA_AUTH_TOKEN": "your-token-here", "CDATA_BASE_URL": "http://localhost:8181/api.rsc", "CDATA_WORKSPACE": "your-workspace-uuid-here", "DISABLE_SSE": "true" } } } }

Server Usage: HTTP Transport (Remote Deployments)

TheHTTP transportis designed for server deployments where the MCP server runs on a remote machine and accepts API requests. This is ideal for:

- Team deployments
- Docker/Kubernetes environments
- Integration with web applications
- Remote access scenarios

- GET /mcp/v1/info- Server and protocol information
- GET /mcp/v1/health- Health check
- POST /mcp/v1/message- Send MCP requests
- GET /mcp/v1/stream- Server-Sent Events for real-time updates

// Connect to the server const client = new MCPStreamableHttpClient('http://your-server:3000/mcp/v1'); await client.connect(); // List available tools const tools = await client.listTools(); // Call a tool const connections = await client.callTool('read_connections', { action: 'list', top: 5 }); // Set up real-time monitoring client.onNotification = (method, params) => { console.log('Notification:', method, params); };

For development and testing, you can run both transports simultaneously:

This is useful for testing both desktop and server scenarios during development.

- read_connections- List, count, get details, or test connections
- write_connections- Create, update, or delete connections
- get_connection_tables- List tables in connection
- get_table_columns- Get table schema information

- read_jobs- List, count, get details, status, history, or logs
- write_jobs- Create, update, or delete jobs
- execute_job- Run a sync job immediately
- cancel_job- Stop running job
- execute_query- Run custom SQL queries

- read_tasks- List, count, or get task details
- write_tasks- Create, update, or delete tasks

- read_transformations- List, count, or get transformation details
- write_transformations- Create, update, or delete transformations

- read_users- List, count, or get user details
- write_users- Create or update users

- read_requests- List, count, or get request log details
- write_requests- Delete request logs

- read_history- List or count execution history records

- read_certificates- List certificates
- write_certificates- Create certificates

- configure_sync_server- Get or update server configuration

All read/write tools use anactionparameter to specify the operation:

{ "tool": "read_connections", "arguments": { "action": "list", "filter": "contains(Name,'prod')", "top": 10 } }

{ "tool": "write_connections", "arguments": { "action": "create", "name": "MyDatabase", "providerName": "System.Data.SqlClient", "connectionString": "Server=localhost;Database=test;" } }

The HTTP transport provides real-time notifications for:

- Tool execution start/completion
- Job execution progress
- Configuration changes
- Error notifications

// Monitor all server events const eventSource = new EventSource('http://localhost:3000/mcp/v1/stream'); eventSource.onmessage = (event) => { const message = JSON.parse(event.data); if (message.method === 'notifications/job_executed') { console.log('Job completed:', message.params); } };

# Start in development mode with both transports npm run dev:both # Start with stdio only npm run dev:stdio # Start with HTTP only npm run dev:http # Type checking npm run typecheck # Linting npm run lint npm run lint:fix # Testing npm test npm run test:watch npm run test:coverage

# Build image docker build -t cdata-sync-mcp-server . # Run with stdio transport docker run -e CDATA_AUTH_TOKEN=your-token cdata-sync-mcp-server # Run with HTTP transport docker run -p 3000:3000 -e MCP_TRANSPORT_MODE=http -e CDATA_AUTH_TOKEN=your-token cdata-sync-mcp-server

# Start with Docker Compose docker-compose up -d cdata-sync-mcp-both

# Deploy to Kubernetes kubectl apply -f k8s/

# Install as systemd service sudo cp cdata-sync-mcp.service /etc/systemd/system/ sudo systemctl enable cdata-sync-mcp sudo systemctl start cdata-sync-mcp

{ "protocol": "Model Context Protocol", "version": "2025-03-26", "transport": "streamable-http", "endpoints": { "message": "http://localhost:3000/mcp/v1/message", "stream": "http://localhost:3000/mcp/v1/stream" } }

{ "status": "healthy", "transport": "streamable-http", "timestamp": "2024-01-15T10:30:00Z", "pendingRequests": 0, "bufferedMessages": 0 }

{ "jsonrpc": "2.0", "id": "1", "method": "tools/call", "params": { "name": "read_connections", "arguments": { "action": "list", "top": 5 } } }

Server-Sent Events stream providing real-time notifications:

data: {"jsonrpc":"2.0","method":"notifications/tool_execution","params":{"tool":"read_connections","timestamp":"2024-01-15T10:30:00Z"}} data: {"jsonrpc":"2.0","method":"notifications/job_executed","params":{"jobName":"TestJob","result":"success","timestamp":"2024-01-15T10:31:00Z"}}

# Run all tests npm test # Run with coverage npm run test:coverage # Watch mode for development npm run test:watch

src/ ├── __tests__/ │ ├── services/ # Service unit tests │ ├── transport/ # Transport tests │ ├── integration/ # Integration tests │ └── utils/ # Utility tests

- Fork the repository
- Create your feature branch (git checkout -b feature/amazing-feature)
- Commit your changes (git commit -m 'Add some amazing feature')
- Push to the branch (git push origin feature/amazing-feature)
- Open a Pull Request

This project is licensed under the MIT License - see theLICENSEfile for details.

- Documentation: Full API documentation available in thedocsdirectory
- Issues: Report bugs and request features viaGitHub Issues
- Discussions: Community support viaCData Community

- Model Context Protocol Specification
- CData Sync Documentation
- Claude Desktop Configuration

MCP servers for Deephaven to orchestrate data workers and power documentation Q&A with LLMs, enabling AI-driven data workflows.