Skip to main content

Quickstart: MCP, API-first

tip

Choose your onboarding path: cloud for instant access, on-premise for full control.

1. Choose Deployment Mode​

  • runMCP: Create an account to manage servers and orchestrate APIs. The community has created many public APIs you can use or you can create your own.
  • chatMCP: Create an account to integrate agents and interact with APIs via chat. You have access to public APIs created by the community or connect your own.
  • No installation required—just sign up and start building.

On-Premise (Advanced/Enterprise)​

  • Start HAPI Servers manually on your infrastructure.
  • Use La Rebelion's bash scripts to set up local DNS (Caddy + API wrapper).
  • Review prerequisites and ensure your team is ready for manual operations.
tip

Need help? Contact us for support or request a demo to get access to the scripts and tools.

2. Explore Example Projects​

  • For cloud, log in and follow the onboarding guides in the webapp.

  • For on-premise:

    • Install the corresponding binary and start the servers.
    • Or download the release from GitHub.
# Check available API Specs
hapi list
# Start the example project
hapi serve example --mcp --headless

3. Access the API​

  • Cloud: Use the provided dashboard and API docs.

  • On-premise: you can either use the CLI or any MCP client (must support HTTP protocol):

    • --headless mode (shown above) disables the REST/Swagger UI and exposes only the MCP endpoint at http://localhost:3000/mcp. Drop --headless if you also want the Swagger UI at http://localhost:3000/swagger.
    • Use any MCP client to connect to the server, such as chatMCP, MCP Inspector, Postman or curl.

More detailed instructions to access and test the MCP tools are available in the Hello World guide.

Remember 🎗️

On-premise setup requires manual server start and DNS configuration. Use provided scripts for Caddy setup or BYO DNS.

Next Steps​

  • Edit the OpenAPI schema to add endpoints.
  • Explore context and memory features in MCP.
  • Deploy to production (see Deployment Guide).

Further Reading​