nhl-go

The NHL MCP Server exposes data from the NHL Go API through the Model Context Protocol (MCP), enabling LLMs and other tools to query hockey data via a standardized model-context interface. The server bundles API client capabilities (teams, players, schedules, standings) and exposes them as MCP endpoints that can be queried by an MCP-enabled agent or inspector. You can use the provided inspector tool to discover available endpoints and then send structured requests to retrieve team rosters, player stats, game schedules, or standings data. This makes it straightforward to incorporate NHL data into prompts, data pipelines, or automated workflows without manually wiring API calls.

To use the server, start the MCP inspector against the nhl-mcp binary as demonstrated in the README. Once running, you can integrate it with your LLM or tooling by sending model-context requests that leverage the NHL data model (e.g., request current standings for a given division, fetch a player’s stats, or pull a team roster for a specific date). The toolchain is designed to support prompt-driven data enrichment, enabling richer, data-backed conversations about hockey stats and schedules.

Prerequisites:

• Go installed (Go 1.23 or newer)
• Git access to clone the repository

Step-by-step installation:

1. Install Go and set up your environment for Go development.
2. Clone the repository: git clone https://github.com/carsonjones/nhl-go.git
3. Build the nhl-mcp binary: cd nhl-go cd mcp && go build -o nhl-mcp
4. Ensure the nhl-mcp binary is accessible in your PATH or reference its path directly when starting the MCP inspector (as shown in the README).
5. Run the MCP inspector against the nhl-mcp binary to explore and use the MCP endpoints.

Note: The README describes the recommended workflow for building and testing locally, including the debug command using npx to inspect the nhl-mcp binary.

Tips and notes:

• The MCP server is built from the nhl-go repository and exposes NHL API data via MCP-compatible endpoints.
• To test locally, you can use the inspector: npx @modelcontextprotocol/inspector ./nhl-mcp
• If you modify or rebuild the nhl-mcp binary, ensure you re-run the inspector against the updated binary to reflect changes.
• There are no required environment variables listed in the README for running the server; you may want to configure your local paths so the inspector can locate ./nhl-mcp.
• The installation steps reference building both the nhl and nhl-mcp binaries when needed, particularly for local development.
• If you plan to deploy, consider packaging approaches (e.g., docker) and environment-specific configuration, though the README does not provide a ready-made dockerfile.