LAION-AI/Open-Assistant — reverse-engineered prompt

Reverse engineered prompt

GitHub

I want a local version of Open Assistant that feels like the public site. It should have the main website, the chat experience, and the data collection side where people can help improve responses by submitting prompts, ranking answers, and labeling content. The chat should support normal back and forth messaging and let users rate replies with a thumbs up or thumbs down.

Please wire up the backend and all required services so I can run the full stack locally with Docker and open it in the browser at localhost. Include login flow that works for local development, including the email magic link inbox if that is part of the setup. If the live chat model needs a separate inference service, make the app ready to connect to it, but at minimum the rest of the stack should boot cleanly and be usable for development.

Keep the setup practical and close to what this repo already expects. If anything is unclear, check the project docs online and make the simplest working version.

Want more depth? Deep Reverse