Connect Gemini CLI to thirds.ai
Give this prompt to your AI agent. The agent does the task for you.
Read https://thirds.ai/docs/mcp/gemini and help me connect Gemini CLI to thirds.ai and check its tools before making a file
This guide uses Gemini CLI. See the Gemini image examples for designs and prompts. The setup does not apply to the Gemini consumer chat app.
Add the server
Create a key on the API keys page. Store it as THIRDS_API_KEY in the environment that starts Gemini CLI. Add this entry to ~/.gemini/settings.json:
{
"mcpServers": {
"thirds": {
"httpUrl": "https://thirds.ai/mcp",
"headers": {
"Authorization": "Bearer ${THIRDS_API_KEY}"
},
"trust": false
}
}
}
Keep the key out of prompts and shared files. Use httpUrl for Streamable HTTP. The url field selects the older SSE transport. Keep tool confirmations enabled.
The Gemini CLI MCP guide owns the connection fields. Its settings guide explains environment variables in JSON values.
Check the connection
Restart Gemini CLI, then run /mcp. Check that thirds connects and lists tools. If it returns 401, check that the process can read THIRDS_API_KEY and that the key is active.
Ask it to list your saved designs before making a file. Read MCP costs and retries before paid work.