hirotomasato/Qwenpi — reverse-engineered prompt

Reverse engineered prompt

GitHub

Build me an enterprise style API gateway for Qwen AI called Qwenpi. It should let me use Qwen through API formats that OpenAI, Anthropic, Gemini, and Responses clients expect, with streaming chat, tool calling, embeddings, image generation, and a beta text to video endpoint.

I need a clean admin dashboard where I can log in with an admin key, see live gateway status, manage model aliases, add or remove Qwen accounts, monitor rate limits and errors, and switch between dark and light mode.

Also build smart account pooling so traffic gets spread across accounts, unhealthy accounts get paused, rate limited accounts recover automatically, and new accounts can be auto registered with temporary mail providers like GuerrillaMail, MoeMail, or TempMail. Include image and video studio pages with aspect ratio choices and generation history.

Make it easy to run with Docker Compose and also locally with a Python backend and React dashboard. Look up current docs online if you need to.

Want more depth? Deep Reverse