🧠 Open Memory Lane

Open-source, Memory-efficient & usable version of Windows Recall — lightweight background capture + timelapse summaries, built for daily use. 🚀

---

✨ What it does

• 📸 Capture desktop screenshots as WebP (small, fast).
• 🔎 Detect meaningful changes using dhash_bits (OpenCV) — fast perceptual hash; ignores tiny cursor flickers.
• 🗄️ Store images + metadata (timestamp, session, wintitle, app name) in an SQLite index.
• 🎞️ Build 15-min detailed clips and a short daily timelapse summary (timelapse speed derived from summary FPS vs detailed FPS).
• 🔁 Robust backlog processing: missed videos/summaries are queued and created on next run.
• 🖱️ System tray app to control / exit the recorder easily.

---

🖥️ Client-side features

• 📜 Scrollable Timeline View

  • Navigate through your day visually with a smooth, zoomable timeline.
  • Jump to exact sessions (15-min detailed clips or daily summaries).
  • Filter by app name, window title, or time range.

• 🔎 Query the Database

  • Built-in query console to search the SQLite metadata:
    • Example: “Show all Chrome windows between 2–4 PM”
    • Example: “List apps I used more than 30 mins today”
  • Fast results with indexed lookups.

• 📤 Export Data as CSV

  • Export metadata (timestamp, app name, window title, etc.) into a clean CSV.
  • Perfect for analytics in Excel / Power BI / Python notebooks.
  • Optional: include file paths to screenshots for external use.

---

🧩 End-to-end workflow

1. Capture: lightweight background screenshots + metadata storage.
2. Process: timelapse clips auto-generated daily.
3. Explore: open the timeline view, scroll through your day.
4. Query: filter/search by apps, titles, or times.
5. Export: save your activity log to CSV for analysis or sharing.

---

⚙️ Quick start

1. 🐣 Create venv & install:

   uv init
   .venv\Scripts\activate
   uv sync

2. 🛠️ Install ffmpeg and make sure it’s on your PATH.

3. ⚙️ Edit .config.yml (paths, fps, intervals) if needed.

4. ▶️ Run the tray app:

   uv run start_app.py

   or run the start.bat

   or build a single exe with PyInstaller and add a shortcut to shell:startup for autostart. ✨

---

📦 Minimal dependencies

• mss, Pillow, numpy, opencv-python, pystray (see pyproject.toml) ✅

---

🎯 Key design choices

• WebP for compact storage. 🗜️
• dhash_bits via OpenCV for quick, low-false-positive change detection (mouse wiggles ignored). 🧠
• SQLite DB for metadata + reliable backlog processing (never lose pending processing). 💾
• FFmpeg concat + per-image durations / fps-based timelapse to create short, watchable daily summaries. 🎬

---

🧰 Use cases

• ⏱️ Productivity tracking — glance at a short visual timeline.
• 🎞️ Video/Recap generation — daily highlights in minutes.
• 💚 Digital wellbeing — understand app usage visually.

---

📝 Notes

• Runs CPU-only (no GPU required). ⚙️
• All data stays local — privacy-first. 🔒
• Want the summary always X minutes? tweak summary_fps in .config.yml. ⚙️
• Want to change the capture interval? tweak capture_interval in .config.yml. ⚙️
• Want to change the WebP quality? tweak webp_quality in .config.yml. ⚙️

---

Sudharshan TK © 2025 — Built for simplicity & privacy. ❤️

---