Summary

Dayflow automatically generates a visual timeline of your macOS workday activities by analyzing screen recordings. It targets users seeking a private, transparent overview of their actual time spent, offering AI-powered summaries and distraction highlights. The primary benefit is a calm, trustworthy record of daily productivity without manual calendar maintenance.

How It Works

The native macOS app captures screen activity at 1 FPS, processing 15-minute intervals. It sends recent footage to an AI provider (Gemini or local models via Ollama/LM Studio) for summarization. Gemini offers a streamlined 2-LLM-call pipeline, while local models involve a more complex 33+ LLM-call process for on-device analysis. Recordings are automatically cleaned up after three days.

Quick Start & Requirements

• Installation: Download Dayflow.dmg from GitHub Releases and grant Screen & System Audio Recording permissions in macOS System Settings. Developers can build from source using Xcode 15+.
• Prerequisites: macOS 13.0+, Xcode 15+. An optional Gemini API key is required for cloud analysis; Ollama or LM Studio must be set up for local analysis.
• Resources: The app is lightweight (~25MB, ~100MB RAM, <1% CPU). Local AI inference can be GPU-intensive.
• Links: GitHub Releases (implied), Gemini API Key: https://ai.google.dev/gemini-api/docs/api-key.

Highlighted Details

• Automatic timeline generation with concise AI summaries and distraction highlights.
• Minimal system impact via 1 FPS screen recording and 15-minute analysis intervals.
• Privacy-first design: user-selectable AI providers (cloud Gemini with BYO key, or local Ollama/LM Studio).
• Automatic storage cleanup removes recordings older than 3 days.
• Native SwiftUI macOS application with auto-updates via Sparkle.

Maintenance & Community

Contributions via Pull Requests are welcomed. A roadmap outlines planned features like an infinitely customizable dashboard and a daily journal. No specific community channels (e.g., Discord, Slack) are listed.

Licensing & Compatibility

Licensed under the MIT License. The software is provided "AS IS" without warranty. Compatible with macOS 13.0+.

Limitations & Caveats

Local AI models may offer lower summarization quality compared to cloud-based Gemini. On-device AI inference is GPU-intensive on Apple Silicon, impacting battery life. Screen recording functionality is dependent on user-granted macOS permissions.