> ## Documentation Index
> Fetch the complete documentation index at: https://docs.rippit.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Gemini CLI

> Connect Rippit to Gemini CLI from the terminal.

## Add the server

```sh theme={null}
gemini mcp add --transport http rippit https://mcp.rippit.com/mcp
```

Or add it directly to `~/.gemini/settings.json` (user scope) or `.gemini/settings.json` (project scope):

```json theme={null}
{
  "mcpServers": {
    "rippit": {
      "httpUrl": "https://mcp.rippit.com/mcp",
      "authProviderType": "dynamic_discovery"
    }
  }
}
```

Use `httpUrl`, not `url`. The `url` key is for SSE, which Rippit doesn't serve. `authProviderType: "dynamic_discovery"` tells Gemini CLI to discover Rippit's OAuth endpoints and register itself, so you don't configure a client ID, secret, or scopes.

## Sign in

```sh theme={null}
/mcp auth rippit
```

Your browser opens to `app.rippit.com`. Sign in and approve the consent screen.

Gemini CLI also detects the `401` on first use and starts this flow automatically. Run `/mcp auth` on its own to list any servers still needing authentication, and re-run `/mcp auth rippit` if a token expires. Sign-in completes through a local callback on port `7777`, so that port needs to be free.

## Using it

Mention Rippit in your prompt so Gemini CLI reaches for the right tools:

```text theme={null}
Use Rippit to show me the top drivers of contact volume last month.
```

## Next steps

* [What Rippit can do for AI assistants](/mcp-server/overview#what-rippit-can-do-for-ai-assistants)
* [Troubleshooting](/troubleshooting)