Posts

@srt54558 @endermanch @uwukko yeah i think it’s designed for that, but it works with OpenAI-compatible APIs too :)

from my experience, API costs are super cheap, esp if you use Gemini [x] flash models, or the new GPT 4.1 nano model.

also, google offer a free API with generous limits if u let them train on

@srt54558 @endermanch @uwukko so i have some local small models running on my Mac mini, and via tailscale I can have its IP be on OpenWebUI, also I have the OpenRouter API in there so I can access all OpenRouter models they have available :)

@srt54558 @endermanch @uwukko well basically, OpenWebUI is just a UI for these models. imagine if you were on chatgpt’s website but the models were jsut not there. once yo’ve installed OWUI, you can just add API URLs which it calls and can have the models from.

@srt54558 @endermanch @uwukko i managed to get it running pretty well on a RPi with 2GB of RAM :)

@nyaathea ok i don’t think it will bc i forgot i have a power button that’s really hard to press so it’s literally impossible to jailbreak lmao

@nyaathea wait what?!

tell me more

can I get it to run on my iPod touch 2 on iOS 4.2.1?? 😭