nipunbatra8 tears out every AI brain but OpenAI

This fork drops its Anthropic and Google model support and bets everything on a single provider.

infrastructurechat-ui

Most Mike forks keep a menu of AI engines behind the chat - you pick which model answers. nipunbatra8 has gone the other way, ripping out support for Anthropic's Claude and Google's Gemini and leaving only OpenAI's models. The model-picker that let users choose has been stripped from the interface too. It's a clear bet: one vendor, one bill, less to maintain.

The catch is that it looks rushed. One of the four OpenAI models still on offer - their newer reasoning model - is wired up in a way that will likely fail at runtime. And the plumbing that was meant to let each user bring their own API key is now dead on arrival, so everything routes through a single shared account key whether the operator intended that or not.

So what Worth a look if you run a legal-AI product on one vendor and want a lean single-provider setup - but check the model wiring before you ship it.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

1 commit from nipunbatra8/legalos, oldest first. Source extracted verbatim from the harvested git log.

SHA Subject Author Date
663c3130 migrate to openai nipunbatra8 2026-05-23 ↗ GitHub

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-563.md from inside the repo you want the changes in.

⬇ Download capture-thread-563.md