// CLI Agent
Use Gemini CLI with llmrelay
Google's official terminal client for Gemini models. Point it at api.llmrelay.dev and keep your existing tool stack working.
Not callable on llmrelay right now
The setup below is correct, but gemini-3.1-pro-preview is not currently served on llmrelay, so requests for it return model_not_found. Upstream still lists this id but returns 503 "no available Gemini accounts" (verified 4x, 2026-07-30). Capacity gap upstream, not our config.
Point this tool at a model from our live lineup instead.
Recommended model
gemini-3.1-pro-preview
Route
openai-compatible
On llmrelay
$1/M · $6/M
Setup
- 1 npm install -g @google/gemini-cli
- 2 Set GEMINI_API_KEY = your llmrelay key.
- 3 Set GEMINI_API_ENDPOINT = https://api.llmrelay.dev/v1beta.
Config
export GEMINI_API_KEY="sk-llmrelay-..." export GEMINI_API_ENDPOINT="https://api.llmrelay.dev/v1beta"
Smoke test
Copy, paste, run. If you get JSON back with a message, you're good.
gemini "say hi"
Gotchas
- !Only Gemini models are available on this route. To use Claude, use one of the other integrations.
You need an llmrelay key first
Sign up in about 30 seconds — free to create an account, no subscription, $10 minimum top-up.