Hi HN,
I built SheepCat-TrackingMyWork originally to solve my own daily struggles with staying organized and remembering what I actually achieved by 5 PM.
I've spent my career heavily focused on C#, SQL, and JavaScript, using traditional, rigid enterprise ticketing systems. The problem is those tools clash with how my brain processes tasks. Dealing with executive dysfunction and context switching meant I needed a tool that actively helps me track my day without requiring me to constantly manage a complex UI. I built this primarily for myself, but I'm sharing it in the hopes that it might help other "neurospicy" folks who find standard tools overwhelming.
How it works:
Instead of a giant, static to-do list, the app runs in the background and gently prompts you every hour: "What task have you done?" You can also manually add tasks in the meantime.
If connected to a local Ollama instance, the app takes those entries, generates an hourly summary, and saves everything to a local CSV. At the end of the day, the local LLM reads through your daily logs and generates a full summary of all tasks and main takeaways. It basically handles your time-tracking and daily stand-up prep automatically.
The Tech & Privacy:
I stepped out of my usual stack and built the core application in Python. Because work logs are highly personal, keeping data private was a strict requirement. The app is containerized with Docker and hooks into an external Ollama setup to run the LLM locally. Not a single byte of data goes to a third-party API.
The License:
I’ve open-sourced it under the GNU AGPLv3 so anyone can use, modify, and self-host it for free. However, I am also offering a paid commercial license for enterprise teams looking to implement neuro-inclusive tools without the network-copyleft restrictions of the AGPL.
I’m currently transitioning from just building this for myself to being an open-source maintainer, and I'd love your feedback:
Does the hourly prompt + CSV logging sound like a good balance between actively tracking work and minimizing cognitive load?
Thoughts on the Python/Docker architecture and how it manages the external Ollama hooks?
For those who have gone the AGPL/Commercial dual-license route, what were your biggest hurdles converting enterprise users?
Website/Docs:
https://chadders13.github.io/SheepCat-TrackingMyWork-Website...
Happy to answer any questions about the code, the Ollama integration, or the design choices!