So, I made this widget for Ollama on Linux because, for some unknown reason, the Ollama team decided not to implement a way to stop a model while it's generating a response.
I was testing DeepSeek R1 14B this morning, and it was running with an amazing 0.3 tokens/second. But when I tried to stop it—oh yeah, you can't! PC started lagging, and now, you get to open a terminal and kill the process manually with sudo systemctl stop ollama, all while your computer is practically frozen.
So, I got pissed off and built this systray app to solve this with one press of a button.
- Clone the repo
- Install the python requirements
pip install -r requirements.txt - Run
install.sh— it’ll create a service to run on startup (You may need to change theExecStartpath to your python installation location, you can check that withwhereis python).
Want to contribute? Go ahead, feel free to improve this mess.
