I built (pulse) a small floating widget for Windows that shows:
• How many ai models I have loaded
• Real tokens/sec (pulled straight from the ollama.log, not estimated)
• A built-in chat window so I don't need the terminal at all
It also has 4 different shapes (circle, squire, hex, diamond) and a cyberpunk-ish cyan glow. You can drag it around and resize it.
It's a Python+ PyQt5 project, ~800 lines, MIT licensed, works on Windows 7/8/10/11.
Mostly sharing it to see if anyone else finds it useful. Curious what other people are using to monitor their local models.
Currently looking for contributors — especially anyone who wants to add GPU graphs , Linux/macOS support, or new themes. Feedback welcome!
join my discord
please join my idea and bring yours to have the best tool for monitoring ai models ))^_^)) , Zainosta ...
Top comments (0)