DEV Community

Cover image for I built a "Gentle" Work Tracker in Python because Stopwatches gave me Anxiety πŸ‘πŸˆ
Andrew Chadwick
Andrew Chadwick

Posted on

I built a "Gentle" Work Tracker in Python because Stopwatches gave me Anxiety πŸ‘πŸˆ

(Shameless Self-Promotion Alert: I built a free, open-source tool to save my own brain, and I think it might help yours too.)
I’ve been a developer for years, and I have a confession: I am terrible at tracking my time.

I suffer from "time blindness." I get into a flow state, look up, and it’s 5:00 PM. I have no idea what I did for the last 8 hours.
For years, I tried to fix this with standard industry tools. But I found that almost every time-tracker on the market shares the same fatal flaw: They ignore Cognitive Ergonomics.

They are either massive, loud enterprise dashboards that drain my executive function, or they are rigid stopwatches that demand I click "Start/Stop" perfectly every time I switch contexts. The anxiety of the ticking clock actually breaks my focus.
So, I opened up my IDE and built the anti-anxiety alternative using Python.

It’s called SheepCat:Track My Work.

The Philosophy: "Gentle" Tracking
I didn't want a tool that acted like a boss standing over my shoulder. I wanted a tool that acted like a helpful assistant.

Here is how the "Gentle" workflow operates:
The Polite Nudge: Instead of a timer you have to babysit, the app gives you a visually quiet, optional prompt once an hour.
The Gentle Part: It is completely dismissible. If you are deep in a complex function, you hit 'Cancel' (or ignore it) and it vanishes. No guilt. No broken streak. It respects your flow state.
Messy Input is Fine: When you do log, you don't need to select from 50 dropdown menus. You just dump your raw brain contents: "Fighting with the SQL query," "Docs are wrong regarding auth."
Local AI Cleans Up the Mess: This is the killer feature. At the end of the day, SheepCat takes those messy, fragmented Python string logs and passes them to a Local LLM (connecting to your local Ollama instance running Qwen2.5 or Llama3).

It synthesizes your chaos into a clean, professional stand-up report.
Privacy: Because it’s local, your proprietary code snippets and internal rants never leave your machine.
Under the Hood (Python & Local AI)
I built this in Python because I wanted it to be accessible and hackable.

GUI: Custom UI that focuses on "Visual Silence"β€”dark modes, muted colors, and no flashing notifications.
Backend: It hooks into Ollama API endpoints to run the summarization locally.
Storage: Flat-file / Local database. No cloud syncing required.

Why I'm Sharing This (The "Shameless" Part)
I built this to scratch my own itch, but I realized that so many other devs especially those with ADHD or Dyslexia are struggling with the same "admin fatigue".

I just released v1.0.2, which includes a standalone installer (so you don't have to mess with pip install if you don't want to) and a new "Glass Purple" theme designed to reduce eye strain.
It is AGPLv3, meaning it is Source-Available and free for personal use.

I would love for you to roast my code, critique the UI, or just tell me if the concept of "Gentle Tracking" resonates with you.

Check it out on GitHub:
SheepCat Track My Work on GitHub

Happy coding (and gentle tracking)!

Top comments (0)