ICON MCP v103

You are deploying an MCP server that provides authenticated access to ICON MCP v103 endpoints via Bearer tokens and supports optional paid access. This server enables AI agents and LLMs to interact with ICON MCP v103 through a standardized, secure interface, with flexible deployment via Docker or local execution.

How to use

Connect to the server using an MCP client and choose your access mode. If you have your own ICON MCP v103 API key, you can access tools for free by authenticating with your key. If you don’t have an API key, you can pay-per-use using the HTTP 402 protocol. In both modes you will discover a set of available tools you can call to interact with ICON MCP v103 endpoints.

How to install

Prerequisites: make sure you have Docker and Docker Compose installed, or have the Python development environment ready with uv and pip available.

Additional setup and usage notes

The server can run in authenticated mode (free) or payment mode (paid access). When running locally, you can start the HTTP API endpoint at http://localhost:8000/mcp. A local development flow commonly uses an environment variable for the internal API key and a runtime command to start the server via uv.

Security and payments

Two operation modes are supported. Authenticated access uses a client-provided API key to proxy calls to ICON MCP v103. Payment mode uses the server’s own internal API key and verifies payments via the D402 workflow. Ensure you protect your API keys and only expose endpoints to trusted clients.

Troubleshooting

If the server doesn’t start, check the runtime logs for errors related to environment variables or missing dependencies. Authentication errors usually indicate an incorrect API key. For payment mode issues, verify the D402 configuration and the payment signature sent by the client.

Available tools

example_tool

Placeholder tool representing ICON MCP v103 API tool access; replace with actual tools when available