wanted to try out deepseek-r1 this weekend and couldn't find a good simple lightweight UI so made one  

pretty impressed by it, I think local open source modals are getting a lot better!

some thoughts (repo here https://github.com/kevoconnell/deepseek-chat):

lightweight UI to me = being able to start talking instantly assuming you have ollama installed.

I saw https://github.com/open-webui/open-webui , but didn't like it because it took a min to startup

founder @explorelabs

don't take my casts seriously please (most of them at least)

I tried having it write code, and it did a ok job, but IMO not as good as other modals I have tried in the past. Please excuse my bad code

repo: https://github.com/kevoconnell/deepseek-code

here is its attempt at producing a PR for a issue I made on github https://github.com/kevoconnell/chess-engine-llm/pull/3/files (i lol'd at this)

TBF, the second example was literally me just playing around with it, I think if more work was done it would be better. 

As time goes on I want to spend more time with OS models. I would love to live in a world one day where I can have my own mini jarvis, I think it will happen

The use cases for local modals seem more clear to me:

a.) as the rest of the world starts to pick up LLMs, most sensitive industries (i.e lawyers, doctors etc) will probably use local modals

b.) it will perform simple tasks that take a couple min max