Gemini API access without the billing circus.

OpenAI-compatible chat, streaming, tools, models, images and videos — routed across the Vertex Cloud Run pool with per-key limits and live usage tracking.

Open admin

459

healthy proxies

models exposed

RPM default

300

RPD default

Generate a key

Free key · IP-bound on first request · default limits: 5 RPM / 300 RPD.

/v1/chat/completions

Drop-in OpenAI shape

Point OpenAI-compatible clients at this base URL and use your rmt-... key.

tools + stream

Function calls work

OpenAI tools are translated into Gemini function declarations and streamed back as tool_calls.

live limits

Usage visibility

Admin shows per-key RPM/RPD, tokens, last model, recent logs and model breakdown.

Quick tests

Click a tab, then copy. Base URL uses the current host automatically.