Chronoid
Automatic time tracking & productivity for macos
101 followers
Automatic time tracking & productivity for macos
101 followers
AI-powered time tracking for macOS with Pomodoro, website blocking, and deep productivity analytics. Privacy-first and fully local.
BatteryBoi (Open-Source)
I've used tracking apps before, and most are overpriced and not particularly smart. This is both detailed, easy to use and very smart in the way it tracks time across apps. I also love the support for all types of AI models (local and cloud).
Also the frequency of updates - you are a madman @vunguyentuan
I’ve been using the app for two days now and got hooked right away. It’s so well executed and really helps me see where my time goes across different projects. The developer is also very responsive to user feedback. Congratulations on the release! 🥳
@marceeelll Thank you so much Marcel!
Disclaimer: I received a free license—lucky me.
Chronoid tracks every app and activity in great detail, and the UI is genuinely beautiful. One suggestion: please add an “Approve all Categories” button for AI suggestions above a 90–99% confidence threshold. The AI proposes many categories, and a bulk-approve would save time for high‑confidence predictions. Thanks for your hard work!
@vunguyentuan Privacy-first time tracking with Pomodoro built in! ⏱️
How does the AI categorize different types of work automatically?
@mskyow Hey Vlad, the app supports two modes:
Rule-based categorization: works very well if you have fixed rules. For example, you can use file names or folder names as indicators of projects.
AI-based for everything else: there are two AI approaches in Chronoid.
- ML classification using fasttext.cc model (local & private). When you have more than 1000 categorized activities, the app will trigger the training process to help categorize new activities.
We prepare a training data set:
- At 10 a.m., visit youtube.com, see Black Friday sales in 2025 video -> distraction
- At 11 a.m., visit github.com, see how to build a web app -> non-distraction
.... ....
Then when new data comes in, the ML model can predict user intent very fast. I would say it works best when the pattern is clear and has around 70% accuracy.
- LLM-based classification: think about using ChatGPT to help categorize your activities to the correct project. You can use your local private LLM run on your machine or use cloud providers like OpenAI, Gemini .... using API
SnapPay
been using chronoid for a bit and it kinda fixed a problem i’d just accepted as permanent: i finally know where my day actually went without babysitting a timer or filling out some boring spreadsheet at night. i just do my work, then crack it open and go “oh, that’s why today felt chaotic.” it’s weirdly satisfying.
@nathan_tran2 Thanks mate!!!!