Documentation
Base URL: https://www.gemini-api.site
Authentication
All /v1/ endpoints require a bearer token, sent as an Authorization header. Tokens are issued privately — there is no public sign-up.
Authorization: Bearer YOUR_KEY
Models
| Model | Description |
|---|---|
gemini-3.8-flash | Default. Fast, current-generation flash model. |
gemini-3.5-flash | Previous generation, kept as a fallback. |
gemini-2.5-flash | Older generation, kept for compatibility. |
model:auto | Let the gateway pick the default model. |
List models
GET /v1/models
Chat completions
OpenAI-compatible chat endpoint.
POST /v1/chat/completions
Content-Type: application/json
{
"model": "model:auto",
"messages": [
{"role": "user", "content": "Hello!"}
]
}
Rate limits
This is a personal gateway. Fair use applies; abusive traffic is blocked at the gateway. If you are seeing 503 responses, Google's side is busy — retry after a short wait.
Changelog
- 2026-11-14 — Default model moved from
gemini-3.6-flashtogemini-3.8-flash. - 2026-11-17 — Gateway moved to new hardware; no API changes.