Nativu5/Gemini-FastAPI

Web-based Gemini models wrapped into an OpenAI-compatible API.

What it solves

Gemini-FastAPI provides a way to access Google's web-based Gemini models via an API without requiring an official Google API key. It wraps these models into an OpenAI-compatible API, allowing users to integrate Gemini into existing tools and applications that expect the OpenAI format.

How it works

The project uses web cookies (__Secure-1PSID and __Secure-1PSIDTS) from a signed-in Google account to authenticate requests to Gemini's web interface. It is powered by the Gemini-API library. The server is built with FastAPI and supports multi-turn conversations using LMDB-based storage for persistence.

Who it’s for

Developers who want to use Gemini's web-based capabilities—including Google Search integration and multi-modal inputs—for free, or those who need an OpenAI-compatible endpoint for their Gemini-powered applications.

Highlights

  • OpenAI Compatibility: Supports /v1/chat/completions and /v1/models endpoints, including streaming, tool calling, and structured JSON output.
  • No API Key Needed: Uses session cookies for authentication, enabling free access to web-based models.
  • Multi-modal Support: Handles text, images, and file uploads.
  • Load Balancing: Distributes requests across multiple Google accounts with individual proxy settings to manage rate limits.
  • Context Management: Offers strategies like compaction (summarizing older turns) or file attachment for handling oversized contexts.
  • Conversation Persistence: Uses LMDB to maintain conversation history across sessions.

Related

  • Project
    zhiyu1998/Gemi2Api-Server

    Gemi2Api‑Server is a self‑hosted FastAPI proxy that forwards OpenAI‑style requests to Google Gemini’s unofficial web API. By supplying your own Gemini login cookies, you get `/v1/models` and `/v1/chat/completions` endpoints, an admin UI, optional model overrides, and an image‑proxy feature—all runnable via Docker or directly with Python.

  • Project
    ntthanh2603/gemini-web-to-api

    A local API server that transforms the Google Gemini web interface into a REST API using browser cookies instead of API keys.

  • Project
    Sophomoresty/gemini-web2api

    gemini‑web2api is a Python server that proxies Google Gemini’s private web API and presents it as an OpenAI‑compatible endpoint (`/v1/chat/completions`, `/v1/models`, etc.). It supports optional API‑key auth, streaming, function calling, image input, multiple Gemini model aliases, adjustable reasoning depth, and can run directly or inside Docker.

  • Project
    Amm1rr/WebAI-to-API

    A browser-native AI runtime that exposes browser-based AI services, such as Google Gemini, through an OpenAI-compatible API layer.

  • Project
    jenslys/opencode-gemini-auth

    A Gemini OAuth plugin for Opencode that enables authentication via Google accounts to access Gemini plans and quotas.