Developer Resources

If you build with an AI coding assistant, you can connect it to the Forage documentation so it answers from our live docs instead of from memory. This page sets up two tools that do that: the Forage MCP (Model Context Protocol) server and llms.txt.

Neither tool is required to integrate with Forage. They make the work faster, so they're worth connecting before you write your first request.

Prerequisites: an AI coding tool that supports MCP, such as Cursor or Claude Code. For llms.txt, any tool that reads a documentation index.

Forage MCP server

The Forage MCP server lets an AI coding assistant look up the Forage API documentation while you work. Your assistant pulls our reference material on demand, so its answers about the Forage API come from the live docs rather than its training data.

Four of its five tools return documentation only, with no access to your account, customer, or benefit data. The fifth, execute-request, runs live API calls, and it works only after you add your own Forage credentials to your MCP client. Those are real requests made as you, and they can reach the live environment rather than the sandbox, so read the Forage MCP Server Tool Reference before you use it.

The server covers our API reference only, not our written guides. For a question about setting up a whole integration, check the guide for that integration as well, since the API reference alone can leave out steps the guide covers. Integration Options points you at the right one.

The server also won't make the code your assistant writes correct, so review what it generates the same way you'd review any AI output.

It's a remote server: just a URL, nothing to install.

https://docs.joinforage.app/mcp

Connect it

When your client asks how to add the server, choose the remote / HTTP / URL option, not a local or command-based one.

Cursor: add the server to .cursor/mcp.json in your project (or ~/.cursor/mcp.json for all projects):

{
  "mcpServers": {
    "forage-docs": {
      "url": "https://docs.joinforage.app/mcp"
    }
  }
}

Claude Code: add it from your terminal:

claude mcp add --transport http forage-docs https://docs.joinforage.app/mcp

Other editors: add the same URL as a remote (HTTP) MCP server in your client's MCP settings. The field names differ, but the value is always the URL above.

Check that it worked

Your client should list forage-docs as a connected MCP server. To test it, ask your assistant to list every Forage API call it can see. A long list of paths means it's connected. An error, or nothing back, means it isn't.

If it's not connecting

• Confirm you added it as a remote (HTTP) server, not a local one. A local entry with a URL fails silently.
• Restart or reload your editor after saving the config.
• Check the URL for stray spaces or trailing characters.

Opening the URL in a browser shows a short notice rather than a page, since it only responds to MCP clients. That's expected.

llms.txt

llms.txt is a plain-text index of our public documentation, published at a standard location that AI agents check:

https://docs.joinforage.app/llms.txt

It lists our guides and API reference as links, so an agent can find the right page and read it on its own. It serves public documentation only, with no account or customer data and nothing to authenticate.

Many AI tools discover llms.txt automatically. If yours lets you add a documentation URL by hand, use the one above.

Next steps

• Read How Forage Works for the model in a couple of minutes.
• Set up Authentication to make your first call.