Gemini API proxy

One Proxy for Gemini and Everything Else

Keep one client config; route Gemini alongside Claude, Codex and the full model list.

1

Create an API key

Sign up and create a key in the console, with an optional budget.

2

Point at /v1

Set your client's base_url to https://apitokenrouter.com/v1.

3

Swap the config

Drop in the new key. Your existing SDK calls stay as they are.

Endpoint (OpenAI-compatible)

https://apitokenrouter.com/v1

Why proxy Gemini at all

Every extra vendor means another key, another balance, another invoice and another client configuration to keep in sync.

Through one OpenAI-compatible endpoint the model family stops being a configuration problem: same key, same base URL, same bill. Switching which model serves a request is a parameter change, not a migration.

What the proxy handles

  • Smart routing — requests are steered across channels by state;
  • Automatic fallback — a failing channel is switched out, retries are the platform's job;
  • Prompt caching — repeated prefixes hit cache instead of being billed twice;
  • One bill — Gemini settles alongside everything else;
  • Consumption tracking — usage per model is visible in the console.

Docs and clients

Integration guides for the API itself and for the clients people actually run are in the docs. Image-model guides are documented there too.

Per-model prices live on the model price list; this page quotes no numbers.

FAQ

Do I need a separate Gemini SDK?

No; the OpenAI-compatible endpoint covers Gemini models on the platform.

Are image models documented?

Yes, Gemini image API guides exist in our docs.

How is usage billed?

One unified bill with consumption tracking in the console; per-model prices are listed on the model price list page (/models.html).

What happens if a channel fails?

The platform performs smart scheduling with automatic fallback across channels.

Where do I see pricing?

On the model price list page (/models.html); this page does not quote prices.

Add Gemini without adding a vendor

One endpoint, one bill — the model list changes, your integration does not.