Hi Product Hunt! 👋 I'm the maker of DoWhat. Let me be honest about why I built this: I was drowning in context switching. Every day I would tackle a dozen different tasks, but when Friday rolled around and it was time to write my weekly summary, my mind went completely blank. I literally couldn't remember what I had accomplished, and my task planning was always in a state of chaos. I needed a "second brain" that could just watch me work and figure it out for me, but I absolutely refused to send my screen data to a cloud server. So, I built DoWhat. It’s an open-source, local-first AI work agent. Here is what it does for you: 📸 **Passive Tracking:** Takes a screenshot every 15s. No manual logging required. 🧠 **AI Summaries:** Uses any OpenAI-compatible vision API (or local models) to group your activities into 15-minute slots with AI-generated summaries. Your weekly reviews write themselves. 🎯 **Proactive Planning:** Automatically extracts actionable tasks from your screen and builds your backlog. 🤖 **OpenClaw Integration:** Connects to IMs (like WeChat/Feishu) and delegates simple tasks to an autonomous agent. **Privacy is the core feature.** Everything (screenshots, SQLite database) stays on your local machine. I built this to solve my own mess, and it’s still early days for the codebase. I would absolutely love your feedback, feature requests, or any architectural critiques! Check out the GitHub repo, and if it saves your Friday afternoons like it did mine, a ⭐️ would mean the world to me. Happy to answer any questions!

DoWhat

Open-source AI digital twin to passively track your workday.

Open-source AI digital twin to passively track your workday.