Skip to content
Back to docs

Install the Goodeye MCP server

Connect Goodeye to your AI agent through the Model Context Protocol. Once connected, the agent can call every Goodeye capability from any conversation: browse and fetch templates, fork them into your account, design new workflows, deploy and run verifiers, manage teams and grants, and more.

Server URLhttps://mcp.goodeyelabs.com/mcp

Claude.ai and Claude Desktop

Claude.ai and Claude Desktop share connectors, so you only set this up once.

  1. Open Settings › Connectors in Claude.
  2. Click Add custom connector.
  3. Enter these values:
    Name
    Goodeye
    URL
    https://mcp.goodeyelabs.com/mcp
  4. Click Add, then sign in with your Goodeye account.
  5. Enable the connector in any chat by clicking the + button.

Claude Code

  1. Register the server from your shell:
Add to Claude Code
claude mcp add --transport http goodeye https://mcp.goodeyelabs.com/mcp
  1. Inside Claude Code, run the slash command:
Authenticate
/mcp
  1. Pick goodeye from the panel and follow the browser flow to sign in with your Goodeye account.

The add command writes to your user-level config, available in every project. Add --scope project to scope it to a single project’s .mcp.json. Use Clear authentication in the /mcp menu to revoke access.

Cursor

Add Goodeye to your project’s .cursor/mcp.json:

Cursor MCP config
{
  "mcpServers": {
    "goodeye": {
      "url": "https://mcp.goodeyelabs.com/mcp"
    }
  }
}

Restart Cursor, then sign in with your Goodeye account when prompted.

ChatGPT

  1. Open Settings › Apps › Advanced settings and enable Developer Mode.
  2. Click Create app.
  3. Enter these values:
    Name
    Goodeye
    MCP Server URL
    https://mcp.goodeyelabs.com/mcp
    Authentication
    OAuth
  4. Confirm and create, then sign in with your Goodeye account.

Developer Mode must stay on. Conversations using custom connectors show an orange Developer Mode indicator.

VS Code (GitHub Copilot)

Requires VS Code 1.99 or newer with GitHub Copilot enabled.

  1. Open the Command Palette (Cmd/Ctrl + Shift + P).
  2. Run MCP: Add Server.
  3. Pick HTTP.
  4. Enter these values:
    URL
    https://mcp.goodeyelabs.com/mcp
    Name
    goodeye

Or add directly to .vscode/settings.json:

VS Code settings
{
  "mcp": {
    "servers": {
      "goodeye": {
        "url": "https://mcp.goodeyelabs.com/mcp"
      }
    }
  }
}

Sign in with your Goodeye account when prompted.

Windsurf

  1. Open Windsurf Settings (Cmd/Ctrl + ,).
  2. Search for MCP, or open Cascade › MCP Servers › View raw config.
  3. Add Goodeye to the mcpServers object:
Windsurf MCP config
{
  "mcpServers": {
    "goodeye": {
      "serverUrl": "https://mcp.goodeyelabs.com/mcp",
      "disabled": false
    }
  }
}

Save, click Refresh (or restart Windsurf), and sign in with your Goodeye account when prompted.

Use an API key instead of OAuth

Skip the browser sign-in flow with a personal API key. Useful for scripted setup or shared dev environments where the browser flow is awkward. Goodeye API keys have the format good_live_<ulid>_<secret> and are shown only once at creation time.

  1. Create a personal key (see the register and sign in docs if you do not have an account yet):
    Create a key
    goodeye auth create-key --name "my-mcp"
  2. Add an Authorization: Bearer good_live_... header in any client config above. Cursor example:
    Cursor MCP config with API key
    {
      "mcpServers": {
        "goodeye": {
          "url": "https://mcp.goodeyelabs.com/mcp",
          "headers": {
            "Authorization": "Bearer good_live_..."
          }
        }
      }
    }

Troubleshooting

401 Unauthorized
Disconnect and reconnect Goodeye in your client. If using an API key, confirm the header reads Authorization: Bearer good_live_....
Tools do not appear
Restart the client after editing config. Confirm the URL is exactly https://mcp.goodeyelabs.com/mcp. In VS Code, agent mode must be enabled.