This project provides a modular FastAPI web server to manage requests for AI services, primarily Gemini, without requiring API keys. It's designed for researchers and developers needing a flexible way to integrate AI models into applications, offering an extendable architecture for easy configuration and switching between providers.

How It Works

The application uses a modular architecture with FastAPI for API endpoints and a service layer for AI model interaction. It leverages browser cookies for authentication with Gemini, automatically retrieving them if not explicitly provided in the configuration. This approach bypasses the need for API keys, simplifying setup and usage for services that support cookie-based authentication.

Quick Start & Requirements

• Install dependencies: poetry install
• Run the server: poetry run python src/run.py
• Configuration: Copy config.conf.example to config.conf and edit.
• Browser: Requires a supported browser (Firefox, Brave, Chrome, Edge, Safari) for cookie retrieval if not provided.
• See: Project Repository

Highlighted Details

• Provides OpenAI-compatible endpoints (/v1/chat/completions).
• Supports Gemini via cookie-based authentication, eliminating API key requirements.
• Modular design allows for easy addition of new AI services.
• Includes endpoints for chat, translation, and general completions.

Maintenance & Community

The project is maintained by Amm1rr. There are no explicit mentions of community channels or a roadmap in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: Suitable for commercial use, but the author notes it's a research project and requires additional security/error handling for production.

Limitations & Caveats

This is a research project and is not recommended for production use without significant enhancements to security and error handling. Gemini is the only currently supported AI service, with plans for others discontinued.