Skip to content

owgit/memento-native

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Memento Native

Your Mac's photographic memory — automatically records your screen and lets you search anything you've seen.

Ever closed a tab and forgot the URL? Lost an important message? Can't remember where you saw that code snippet? Memento captures your screen continuously, extracts all text using OCR, and makes it searchable. Everything stays 100% local on your Mac.

Swift macOS License Buy Me A Coffee

Native Swift rewrite of apirrone/Memento. Open-source alternative to Rewind.ai.

Use Cases

  • 🔍 Find by keyword — Search "invoice", "meeting", "password"
  • 💬 Recover lost text — Find that message or email you closed
  • 🐛 Debug timeline — Scroll back to see what happened
  • 🧠 Semantic search — Find "coding tutorial" even if text says "programming lesson"
  • 📋 Visual history — Browse your screen activity by time

Features

Feature Description
📸 Screen Recording ScreenCaptureKit (macOS 14+)
🔍 OCR Search Apple Vision text recognition
🧠 Semantic Search Find by meaning, not just keywords
🎨 App Learning Auto color-codes apps in timeline
📹 H.264 Video Hardware-accelerated encoding
💾 Full-Text Search SQLite FTS5
⚡ Lightweight ~1% RAM, minimal CPU
🔒 Privacy-First No cloud, no telemetry

Installation

Option 1: DMG (Recommended)

Download the latest DMG from Releases, open it, and drag both apps to Applications.

Option 2: Build from source

git clone https://github.com/owgit/memento-native.git
cd memento-native
./build-dmg.sh        # Creates dist/Memento-Native-1.0.0.dmg
# Or build individually:
cd MementoCapture && ./bundle.sh
cd ../MementoTimeline && ./bundle.sh

Setup Permissions

  1. Open System SettingsPrivacy & SecurityScreen Recording
  2. Click + and add /Applications/Memento Capture.app
  3. Enable it (toggle ON)
  4. Start the app: open /Applications/Memento\ Capture.app

Requirements

Requirement Details
macOS 14.0 Sonoma or later
Mac Apple Silicon (M1/M2/M3) or Intel Mac
RAM 8GB minimum
Disk ~150MB/hour of recording
Permission Screen Recording

Performance

Metric Value
CPU ~1-3% (idle between captures)
RAM ~50-100MB
Capture interval Every 2 seconds
Video codec H.264 hardware-accelerated

How It Works

┌─────────────────┐     ┌──────────────┐
│ MementoCapture  │────▶│   SQLite     │◀────│ MementoTimeline │
│                 │     │  + FTS5      │     │                 │
│ • Screenshot    │     │  + Vectors   │     │ • View frames   │
│ • Vision OCR    │     └──────────────┘     │ • Text search   │
│ • H.264 encode  │            │             │ • Semantic search│
│ • Embeddings    │            ▼             └─────────────────┘
└─────────────────┘     ~/.cache/memento/

Semantic Search

Uses Apple NaturalLanguage for on-device embeddings:

// 512-dim sentence embedding → Int8 quantized (8x compression)
NLEmbedding.sentenceEmbedding(for: .english)

Data stored in ~/.cache/memento/

Privacy

  • 100% offline — works without internet
  • No accounts — no sign-up required
  • No telemetry — zero data collection
  • Clipboard opt-in — clipboard monitoring disabled by default, toggle in menu
  • Local storage — delete anytime with rm -rf ~/.cache/memento

Alternatives

App Platform Price Privacy
Rewind.ai macOS $19/mo Cloud
Memento (Python) Linux Free Local
Memento Native macOS Free Local

Roadmap

🤖 AI-Powered Search (Coming)

Local LLM integration for natural language queries:

  • "What was that article about React I read yesterday?"
  • "Find the Slack message from Johan about the API"
  • "Show me when I was working on the login bug"
  • "What did I copy to clipboard around 3pm?"

License

MIT — use it however you want.


Keywords: screen recorder macos, ocr search mac, rewind alternative, local screen recording, privacy screen capture, searchable screenshots, macos productivity, swift screencapturekit

About

Open-source Rewind.ai alternative for macOS. Local screen recording with OCR search.

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published