This project provides a personal, local work log recorder that automatically captures application usage, website browsing, screen content, and time spent. It targets individuals seeking to understand and review their work habits without manual input, offering a searchable history and optional AI-powered insights, all while prioritizing local data privacy.

How It Works

The system operates by continuously monitoring foreground application changes, browser URLs, and capturing periodic screen content with OCR. It employs idle detection using keyboard/mouse and screen activity to differentiate work time from downtime. Captured data is aggregated into sessions and stored locally in an SQLite database, enabling historical playback and analysis. This approach focuses on creating a personal work "trace" rather than a traditional monitoring tool.

Quick Start & Requirements

• Primary install: Download the latest release (.dmg for macOS, .exe for Windows) from the project's Releases page.
• Non-default prerequisites:
  • macOS: May require running sudo xattr -rd com.apple.quarantine "/Applications/Work Review.app" if the application is flagged as damaged. Screen Recording and Accessibility permissions are needed for the "Desktop Avatar" feature.
  • Windows: .exe installer.
  • Development: Node.js 18+, Rust stable, Tauri 2 CLI.
• Links: CHANGELOG.md, docs/WINDOWS_OCR.md.

Highlighted Details

• Core Tracking: Automatically records application usage (duration, window titles), browser activity (URLs aggregated by site/page), and screen content via timed screenshots with OCR.
• Smart Analysis: Features a "Work Assistant" for natural language queries about recorded activities, time range recognition, session aggregation, to-do extraction from context, and structured daily report generation.
• Desktop Avatar (Beta): An optional, lightweight desktop companion that visually reflects the user's current work status (e.g., working, meeting, idle) with customizable appearance.
• Privacy Controls: Granular settings include per-app tracking levels (normal, anonymized, ignore), sensitive keyword filtering, domain blacklisting, and manual/automatic pausing.
• AI Integration: Supports local (Ollama) or cloud-based AI models for enhanced summarization and Q&A, with core functionality available even when AI is disabled.

Maintenance & Community

The project mentions linux.do community for discussion and support. Information regarding specific contributors, sponsorships, or a public roadmap is not detailed in the provided text.

Licensing & Compatibility

The project is released under the MIT license, which is permissive and generally compatible with commercial use and linking within closed-source projects. All data is stored locally, enhancing privacy and compatibility with various user workflows.

Limitations & Caveats

The "Desktop Avatar" feature is currently in Beta and requires additional macOS permissions (Screen Recording, Accessibility). Initial macOS installations may require manual quarantine attribute removal. The AI-enhanced features are optional and necessitate the user to configure their own AI model services.