About this project

This project is a proxy server that enables interaction with GitHub Models API models while ensuring full compatibility with the standard OpenAI API. It provides access to dozens of models from various providers (including OpenAI, Meta, Microsoft, Mistral, DeepSeek, and xAI). ### Key Features: - **Multiple Interfaces**: Support for the standard OpenAI API, an extended JSON API, a simplified Chat API, and an interactive Web UI for testing. - **Monitoring and Analytics**: A built-in Prometheus-format endpoint for tracking HTTP requests, errors, memory usage, and performance. A separate monitoring web interface is available at `/monitor`. - **Load Control**: Rate limiting mechanisms and concurrency control with a waiting queue to ensure stability. - **Auxiliary Tools**: - `watchdog.sh` for automatic server recovery. - Scripts for benchmarking, model testing, and code example generation. - TTS function (Ukrainian voiceover) with streaming support. - **Intelligent Functions**: Automatic context limit checks and a model recommendation system based on the task (speed, quality, text volume). ### Technical Stack: - Node.js 18+ - Integration with GitHub Models API