multi-model AI infrastructure for prompting and comparing queries to multiple models.

Built with Go backend, vanilla JavaScript frontend, and AI models across isolated VMs. Features side-by-side model comparison, real-time token streaming, and modern space-themed UI.

You will first have to setup your local running models, I used llama.cpp for this, you can find a .service file here for an example of the setup I use.

1. Clone

git clone https://github.com/mengdotzip/ai.meng.zip
cd ai.meng.zip

2. Start backend

cd backend && go run main.go

3. Serve frontend

cd frontend

Then serve the files with something like nginx (or use https://github.com/mengdotzip/Mazarin If you are sigma)

4. Open http://localhost:8080

Don't forget to setup the backend to point to your own models, then after that everything should work.