About this project

Dayflow is a private, automatic work journal and time tracker for macOS. Instead of requiring timers or manual notes, it captures lightweight screen activity, analyzes it with an AI provider of your choice, and turns the day into a chronological timeline of activity cards. Key capabilities described in the README: - Automatic timeline: captures screen chunks and converts them into activity cards, so you get a work journal without starting timers. - Context-aware summaries: analyzes what was actually happening on screen, not just which app was active, so tools like Cursor, Chrome, YouTube or Slack become meaningful work context. - Daily standup: pulls yesterday's highlights, today's tasks and blockers from the timeline to help write updates quickly. - Chat with your work journal: ask natural-language questions about your day, week or year and get answers grounded in the timeline. - Weekly review: aggregates the timeline into focus patterns, categories, app usage and interaction graphs. - Distraction tracking: identifies distracting sessions and shows them alongside focused work. - Timeline export: exports the timeline as Markdown for any date range, useful for status updates, client notes or personal reviews. - Local-first storage: recordings, timeline data and the app database stay on the Mac by default, under ~/Library/Application Support/Dayflow/. - AI provider choice: local models via Ollama or LM Studio, Gemini with your own API key, or ChatGPT/Claude through local CLI tools. Cloud providers receive the activity data needed for analysis; local models keep analysis on the machine. - Automatic cleanup: configurable storage limits let Dayflow purge old recordings automatically. Installation options include downloading the latest Dayflow.dmg from GitHub Releases or using Homebrew (brew install --cask dayflow). Requirements are macOS 14+ and Screen & System Audio Recording permission, plus an optional AI provider such as a Gemini API key, Ollama, LM Studio, Codex CLI or Claude Code. The project can be built from source by cloning the repository and opening Dayflow/Dayflow.xcodeproj in Xcode. It is MIT licensed, and issues and pull requests are welcome.