Connect an AI tool

Connect a compatible AI tool to your LFG List account using a Personal Access Token.

Requirements

• LFG Pro for the account creating and using the token.
• A Personal Access Token.
• An MCP client that supports remote Streamable HTTP and a Bearer authorization header.

Connection details

Enter these values in your MCP client’s server configuration:

Setting	Value
Name	LFG List
URL	https://mcp.lfglist.com/mcp
Transport	Streamable HTTP
Authorization header	Authorization: Bearer <your-token>

Store the token in the client’s supported secret or local configuration mechanism. Do not publish it in a repository, screenshot, shared conversation, or bug report. The token is separate from your Google or email/password sign-in.

Generic JSON configuration

If your client supports this configuration shape, use:

{
  "mcpServers": {
    "lfglist": {
      "url": "https://mcp.lfglist.com/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_LFG_LIST_TOKEN>"
      }
    }
  }
}

Client configuration formats differ. Use the endpoint, transport, and header above with the fields your client expects; this JSON is not universal.

Agent setup brief

You can give your agent this setup brief without putting the real token in the conversation:

Connect to the hosted LFG List MCP server at https://mcp.lfglist.com/mcp.
Use remote Streamable HTTP and Authorization: Bearer <token>.
Help me put my Personal Access Token in this client's trusted local secret configuration.
Do not print the token or put it in tracked files.
After connecting, call list_projects to verify read access.
Do not change project or item data during the connection test.

Smoke test

1. Save the configuration and reconnect the MCP client as required by that client.
2. Ask it to call list_projects.
3. Confirm the projects belong to the expected LFG account.
4. Select a project by its returned ID or unambiguous name before requesting further work.

A successful connection does not grant extra project permissions. Shared access still depends on membership and Pro requirements. The hosted service should not need your Firebase login or browser session.

If the test fails, use MCP troubleshooting. For supported actions and replacement semantics, read the tools reference.