Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/JerryZLiu/Dayflow/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

Dayflow’s automatic timeline is the core feature that watches your screen and generates a detailed, AI-powered record of how you spend your time. Unlike traditional time trackers that only log which app is open, Dayflow understands context — the difference between “researching on YouTube” and “watching cat videos.”
Timeline generation runs entirely in the background. You don’t need to manually start or stop tracking — just enable recording and Dayflow handles the rest.

​
How It Works

​
1. Capture

Dayflow takes periodic screenshots of your active display every 10 seconds (configurable). These screenshots are:
  • Scaled to ~1080p resolution for efficiency
  • Compressed as JPEG at 85% quality
  • Stored locally in ~/Library/Application Support/Dayflow/recordings/
  • Never uploaded until you choose to analyze them
Privacy first: Screenshots stay on your Mac. Nothing leaves your machine until analysis is triggered, and only to the AI provider you’ve configured.
The capture system automatically pauses during:
  • System sleep
  • Screen lock
  • Screensaver activation
And automatically resumes when you return.

​
2. Analyze

Every minute, Dayflow checks for unprocessed screenshots and groups them into batches based on:
  • Duration: Batches target your configured duration (varies by AI provider)
  • Continuity: Gaps larger than the configured threshold create new batches
  • Minimum activity: Only batches with 5+ minutes of activity are processed
Batches are sent to your chosen AI provider for analysis.

​
3. Generate Timeline Cards

Your AI provider processes the screenshots and generates timeline cards — structured summaries of your activity: 
{
  "title": "Coding Swift animation work",
  "startTimestamp": "2:15 PM",
  "endTimestamp": "3:45 PM",
  "category": "Development",
  "summary": "Dialing in spring curves and reverse logic for timeline card animations"
}
Each card includes:
  • Time range: When the activity occurred
  • Title: Concise description of what you were doing
  • Summary: AI-generated context about the work
  • Category: Automatically classified (e.g., Development, Communication, Research)
  • Detailed summary: Deeper insights when available

​
4. Display

Timeline cards appear in the main Dayflow UI, organized chronologically. You can:
  • Scroll through your day — see exactly where your time went
  • Review summaries — understand context at a glance
  • Navigate by date — explore past days
  • View app icons — see which applications were active (optional)

​
Context-Aware Tracking

Dayflow’s AI analysis goes beyond simple app monitoring:
Traditional Time TrackerDayflow
”2 hours in Chrome""Researching SwiftUI animation APIs"
"1 hour in Slack""Planning Q1 roadmap with design team"
"30 minutes in YouTube""Learning about Combine framework patterns”
The AI understands:
  • What you were working on (not just which app)
  • Context from window titles, visible text, and UI elements
  • Transitions between different types of work
  • Focus periods vs. context switching

​
Viewing Today

The timeline defaults to “today” using a 4AM boundary — meaning your “day” runs from 4AM to 3:59AM the next morning. This ensures late-night work sessions aren’t artificially split across calendar days.

​
Browsing Past Days

Use the date picker or navigation arrows to explore previous days. Timeline cards are permanently stored in your local database (chunks.sqlite).

​
Searching & Filtering

While Dayflow doesn’t currently have built-in search, you can:
  1. Export your timeline as Markdown (see Export)
  2. Use your favorite text editor to search exported data

​
Storage Management

Screenshots accumulate quickly. Dayflow includes automatic storage cleanup:
  • Configurable limits: Set storage caps from 1GB to 20GB (or unlimited)
  • Automatic purging: Old screenshots are deleted when limits are reached
  • Manual cleanup: Quit Dayflow and delete ~/Library/Application Support/Dayflow/ to start fresh
See Storage Configuration for details.

​
Performance

Dayflow is designed to be lightweight:
  • App size: ~25MB
  • RAM usage: ~100MB
  • CPU usage: Less than 1% (during capture)
  • Battery impact: Minimal — screenshot capture is far less intensive than continuous video recording

​
Troubleshooting

  1. Check that Screen & System Audio Recording permission is enabled in System Settings → Privacy & Security
  2. Verify your AI provider is configured in Settings → AI Provider
  3. Ensure you have at least 5 minutes of continuous activity
  4. Check the Debug view (menu bar icon → Debug) for batch processing status
  1. Local models: May struggle with complex contexts — consider switching to Gemini or ChatGPT/Claude
  2. Low activity: Short bursts of activity (less than 5 minutes) are skipped
  3. Provider issues: Check your API key or CLI tool authentication
Recording automatically pauses during:
  • System sleep or screen lock
  • Screensaver activation
This is intentional to avoid capturing sensitive content or blank screens.

​
Next Steps

AI Providers

Learn about Gemini, local models, and ChatGPT/Claude options

Export Timeline

Export your timeline as Markdown for external use

Timelapse

Watch visual timelapses of your screen activity

Journal

Track intentions and reflections alongside your timeline