Back to browse
Didon – AI workday reports for productivity analysis

Didon – AI workday reports for productivity analysis

by babakzy·Jun 21, 2026·1 point·0 comments

AI Analysis

●●SolidSolve My ProblemDark Horse

Local LLM analyzes screenshots on-device for privacy-focused time tracking.

Strengths
  • Qwen-3-VL:2b runs locally—screenshots never leave the machine for privacy.
  • Automatic screenshot capture eliminates manual timer start/stop friction.
  • CSV export supports client billing and internal reporting workflows.
Weaknesses
  • macOS only—Windows and Linux users completely excluded from launch.
  • Screenshot-based tracking already exists in RescueTime and ActivityWatch.
Category
Target Audience

Freelancers, indie developers, and remote workers tracking billable hours

Similar To

RescueTime · ActivityWatch · Timely

Post Description

As an indie engineer, I wanted some real feedback on my productivity, like having an actual external boss. I tried using time trackers (even a physical timer), but they weren't good at tracking everything. So I built Didon, an AI time tracker that watches the screen periodically and generates work logs, then summarizes them into your workday report.

Similar Projects

AI/ML●●Solid

I run a vision model on every screenshot, locally, on a 4GB GPU

Runs multimodal screen memory on 4GB VRAM while Microsoft Recall requires high-end NPU.

Dark HorseSolve My Problem
skye0110
3677d ago