IntegrationsDev tool
Gemini CLI
Point Gemini CLI at the RouteMux Vertex-compatible endpoint.
RouteMux exposes a Google GenAI / Vertex-compatible surface, so Gemini CLI can use it with a base URL and key override.
Configure
export GOOGLE_GEMINI_BASE_URL="https://api.routemux.com/vertex-ai"
export GEMINI_API_KEY="sk-..." # your RouteMux keyRouteMux accepts the key in Authorization: Bearer, in x-api-key, or in Google's
x-goog-api-key — whichever your client sends.
Model availability
Vertex-family models are enabled per key. If GET /v1/models does not list the model you
want, calling it will fail regardless of how the client is configured.
Endpoints
| Operation | Path |
|---|---|
| Generate | /vertex-ai/v1beta/models/{model}:generateContent |
| Stream | /vertex-ai/v1beta/models/{model}:streamGenerateContent |
The model segment is URL-encoded, so slugs containing / work unchanged.
Errors
Vertex-family endpoints return Google's error envelope — { error: { code, status, message } } —
not the OpenAI shape. A 402 with status: INSUFFICIENT_CREDITS means an empty wallet.