Skip to main content

Listen

Starts an MCP (Model Context Protocol) server that listens for incoming client connections via Server-Sent Events (SSE). This is the core node for building MCP servers that expose tools, prompts, and resources to AI applications.

Common Properties​

  • Name - The custom name of the node.
  • Color - The custom color of the node.
  • Delay Before (sec) - Waits in seconds before executing the node.
  • Delay After (sec) - Waits in seconds after executing node.
  • Continue On Error - Automation will continue regardless of any error. The default value is false.
info

If the ContinueOnError property is true, no error is caught when the project is executed, even if a Catch node is used.

Options​

  • SSE Server IP - IP address to bind the SSE server to. Default: 127.0.0.1. Use 0.0.0.0 to accept connections from any network interface.
  • SSE Server Port - Port number for the SSE server. Default: 8888. Ensure this port is available and not blocked by firewall.
  • SSE Server Endpoint - URL endpoint path for the MCP server. Default: /sse. Clients connect to http://{IP}:{Port}{Endpoint}.
  • Server Name - Human-readable name for your MCP server. Default: My MCP Server. This appears in client discovery.
  • Server Version - Version string for your MCP server. Default: 0.0.1. Use semantic versioning (e.g., 1.2.3).

Output Ports​

The Listen node has three output ports for connecting different types of MCP capabilities:

  • tools (Port 0) - Connect Tool In nodes here to expose executable tools to clients. Only accepts Tool In nodes.
  • prompts (Port 1) - Connect Prompt In nodes here to expose prompt templates to clients. Only accepts Prompt In nodes.
  • resources (Port 2) - Connect Resource In nodes here to expose resources to clients. Only accepts Resource In nodes.

How It Works​

The Listen node creates an MCP server and registers all connected capabilities when the flow starts:

  1. Initialization - When the flow starts, the node waits for the runtime to be ready
  2. Registration - Scans all connected Tool In, Prompt In, and Resource In nodes on the respective ports
  3. Server Setup - Registers each capability with the MCP server using their configurations
  4. SSE Server - Starts an HTTP server that provides SSE transport for MCP protocol
  5. Request Routing - Incoming requests are routed to the appropriate Tool In/Prompt In/Resource In node
  6. Response Handling - Results from Tool Out/Prompt Out/Resource Out nodes are sent back to clients

Server Capabilities​

The Listen node automatically enables the following MCP capabilities:

  • Tool Support - Clients can list and call tools
  • Prompt Support - Clients can list and retrieve prompts
  • Resource Support - Clients can list and read resources (with subscription support)
  • Logging - Server-side logging for debugging

Requirements​

  • Network Access - The specified port must be available and accessible
  • Connected Nodes - At least one Tool In, Prompt In, or Resource In node should be connected
  • Firewall - Ensure firewall rules allow incoming connections on the configured port

Usage Notes​

Server Lifecycle​

  • The MCP server starts automatically when the flow begins execution
  • The server runs continuously until the flow is stopped or encounters an error
  • Proper cleanup is performed when the flow stops (via OnClose)

Connection Management​

  • Each connected Tool In/Prompt In/Resource In node is registered at startup
  • If a node registration fails, a warning is logged but the server continues
  • Client connections are managed by the SSE transport layer
  • The server supports multiple concurrent client connections

Network Configuration​

  • Local Testing - Use 127.0.0.1 to restrict access to localhost only
  • Network Access - Use 0.0.0.0 to allow connections from other machines
  • Port Selection - Choose a port that doesn't conflict with other services
  • Endpoint Path - The default /sse endpoint works for most cases

Error Handling​

The node handles errors gracefully:

  • Port Already in Use - Server will fail to start if port is occupied
  • Registration Failures - Individual node registration failures are logged as warnings
  • Runtime Not Ready - Waits with retry logic until runtime is available
  • Context Cancellation - Properly shuts down when flow is stopped

Security Considerations​

  • The SSE server does not include authentication by default
  • Consider using a reverse proxy (nginx, Caddy) for production deployments
  • Use HTTPS in production to encrypt communication
  • Restrict IP binding to 127.0.0.1 if only local access is needed
  • Implement network-level security (firewall, VPN) for sensitive deployments

Examples​

Example 1: Basic MCP Server with Tools​

Setup:

  1. Add a Listen node
  2. Configure:
    • Server Name: "My Automation Server"
    • SSE Server Port: 8888
  3. Connect a Tool In node to the tools port
  4. Connect a Tool Out node after your automation logic

Result: An MCP server that exposes one or more automation tools to AI clients.

Example 2: Multi-Capability Server​

Setup:

  1. Add a Listen node
  2. Connect multiple nodes:
    • 3 Tool In nodes → tools port
    • 2 Prompt In nodes → prompts port
    • 2 Resource In nodes → resources port
  3. Configure each capability with appropriate settings

Result: A comprehensive MCP server offering tools, prompts, and resources.

Example 3: Network-Accessible Server​

Setup:

  1. Add a Listen node
  2. Configure:
    • SSE Server IP: 0.0.0.0
    • SSE Server Port: 9000
    • Server Name: "Company Automation Hub"
    • Server Version: "1.0.0"

Connection URL: http://your-server-ip:9000/sse

Result: An MCP server accessible from other machines on the network.

Example 4: Custom Endpoint Path​

Setup:

  1. Add a Listen node
  2. Configure:
    • SSE Server Endpoint: /mcp/api
    • SSE Server Port: 8080

Connection URL: http://127.0.0.1:8080/mcp/api

Result: Server with a custom endpoint path for organizational purposes.

Best Practices​

  1. Server Configuration:

    • Use descriptive server names that identify the purpose
    • Follow semantic versioning for Server Version
    • Document the capabilities your server provides

  2. Network Setup:

    • Start with 127.0.0.1 during development
    • Use specific IP addresses in production, not 0.0.0.0
    • Choose high ports (8000+) to avoid privileged port issues

  3. Capability Organization:

    • Group related tools, prompts, and resources logically
    • Use clear, descriptive names for each capability
    • Document the expected inputs and outputs

  4. Production Deployment:

    • Use a reverse proxy with HTTPS for production
    • Implement proper authentication at the proxy level
    • Monitor server logs for errors and performance
    • Set up health checks and monitoring

  5. Testing:

    • Test with local connections first
    • Verify all capabilities are registered correctly
    • Test error handling scenarios
    • Validate with multiple concurrent clients

  6. Resource Management:

    • Ensure long-running servers don't leak resources
    • Monitor memory usage for servers with many capabilities
    • Implement proper cleanup in connected nodes

Common Errors and Solutions​

Error: "Port already in use"​

Cause: Another service is using the configured port.

Solution:

  • Change the SSE Server Port to an available port
  • Stop the conflicting service
  • Use netstat -ano | grep :8888 (Linux/Mac) or netstat -ano | findstr :8888 (Windows) to find the process

Error: "Flow stopped before starting SSE server"​

Cause: Flow was stopped during server initialization.

Solution:

  • This is expected when stopping the flow during startup
  • Ensure the flow runs long enough for initialization

Warning: "Failed to register tool/prompt/resource"​

Cause: Configuration error in a connected node.

Solution:

  • Check the configuration of the referenced Tool In/Prompt In/Resource In node
  • Ensure all required fields are filled
  • Verify the node is properly connected to the correct port

Tips for Effective Use​

  1. Start Simple - Begin with one capability type, then expand
  2. Use Meaningful Names - Server name and version help clients identify your server
  3. Test Locally First - Use localhost binding during development
  4. Monitor Logs - Watch for registration warnings and connection issues
  5. Document Your Server - Maintain documentation of available capabilities
  6. Version Your Server - Increment version when capabilities change
  7. Plan for Scale - Consider load if many clients will connect
  8. Secure Production - Always use HTTPS and authentication in production