Skip to main content
Gemini CLI is Google’s open source agent for the terminal. It can call Spyro AI tools while it works on your project.

Connect with sign-in

1

Add the server

Add it to ~/.gemini/settings.json (or .gemini/settings.json for one project). The httpUrl field selects the streamable HTTP transport:
~/.gemini/settings.json
Or use the command: gemini mcp add --transport http spyro-ai https://mcp.spyro.app/mcp
2

Sign in

Start Gemini CLI and run:
Sign in in your browser, pick your workspaces and click Allow.

Connect with a key

Use a key for headless or CI use. Create it in Settings → MCP.
~/.gemini/settings.json
Set SPYRO_AI_MCP_KEY in your environment first. If your version does not expand the variable, check Gemini CLI’s configuration guide, and keep the key out of any file you commit.

Limit which tools it can use

Gemini CLI can allow or block individual tools per server. This is useful to keep it away from the tools that make changes:
A key or approval with Read access already hides these tools. Excluding them is a second layer.

Manage it

Inside Gemini CLI, /mcp lists servers, their status and the tools they offer. Use it to confirm the Spyro AI tools loaded.

Try these prompts

  • “Use Spyro AI to list my critical audit issues and fix the ones in this repo.”
  • “Ask Spyro AI which pages lost the most clicks in the last 28 days.”
  • “Audit https://example.com/blog/post with Spyro AI and rewrite the intro.”

If it fails

  • Server not listed in /mcp: check the JSON is valid and the key is httpUrl, not url (which means the older SSE transport).
  • 401 error: sign in again with /mcp auth spyro-ai, or check the key.
  • More: Troubleshooting.
Google’s guide: MCP servers with Gemini CLI.