Meero is a privacy-first, cross-platform activity logging system that tracks your computer usage with surgical precision. It captures window activity, browser URLs, clipboard content, screenshots with OCR, and advanced behavioral metrics—all stored locally in a SQLite database.
⚡ Use Case: Productivity tracking, time auditing, research data collection, or building context-aware AI assistants.
| Dashboard (Coming Soon) | Metrics View |
|---|---|
| [Add screenshot here] | [Add screenshot here] |
- Active Window Tracking – Logs application name and window title
- Browser URL Extraction – OCR-based address bar reading (no clipboard interference)
- Clipboard Monitoring – Tracks text changes with privacy filters
- Idle Detection – Platform-specific (Windows/macOS/Linux)
- Session Management – Active vs idle time tracking
- Screenshot Capture – Periodic full-screen captures
- OCR Processing – Extracts text using Tesseract
- Smart Retention Policies – Keep all, delete after OCR, or keep only meaningful content
- ⌨️ Typing Speed – Keys per minute with app-level breakdown
⚠️ Error Detection – Identifies errors in window titles- 🔄 Context Switching – Measures app switching frequency
- 😴 Idle Recovery – Time to refocus after breaks
- 📑 Tab Explosion – Detects rapid browser tab switching
- 📈 Productivity Score – Hourly aggregated metrics
- Pause/Resume Logging – Toggle tracking anytime
- Ignore List – Exclude specific applications
- Sensitive Word Filtering – Redacts passwords, emails, credit cards
- 100% Local Storage – No data leaves your machine
┌─────────────────────────────────────────────────────────────┐
│ MEERO LOGGING SYSTEM │
├─────────────────────────────────────────────────────────────┤
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────────────┐ │
│ │ Activity │ │ Browser │ │ Screen │ │
│ │ Tracker │ │ Tracker │ │ Tracker │ │
│ │ │ │ (OCR) │ │ (Screenshot/OCR) │ │
│ └──────┬──────┘ └──────┬──────┘ └──────────┬──────────┘ │
│ │ │ │ │
│ ▼ ▼ ▼ │
│ ┌──────────────────────────────────────────────────────┐ │
│ │ CORE SCHEDULER │ │
│ │ (Multi-threaded, WAL Mode SQLite) │ │
│ └──────────────────────────────────────────────────────┘ │
│ │ │ │ │
│ ▼ ▼ ▼ │
│ ┌─────────────┐ ┌─────────────┐ ┌─────────────────────┐ │
│ │ Clipboard │ │ Keyboard │ │ Metrics │ │
│ │ Monitor │ │ Tracker │ │ Calculator │ │
│ └─────────────┘ └─────────────┘ └─────────────────────┘ │
│ │ │
│ ▼ │
│ ┌──────────────────────────────────┐ │
│ │ SQLite Database │ │
│ │ ~/.personal_ai_logger/ │ │
│ │ activity.db │ │
│ └──────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────┘
meero/
├── logger/ # Core logging system
│ ├── core/
│ │ ├── scheduler.py # Multi-threaded task scheduler
│ │ ├── database.py # SQLite with WAL mode
│ │ └── logger.py # Application logging
│ ├── trackers/
│ │ ├── activity.py # Window tracking
│ │ ├── browser.py # OCR-based URL extraction
│ │ ├── screen.py # Screenshot + OCR
│ │ ├── clipboard.py # Clipboard monitoring
│ │ └── keyboard.py # Typing speed metrics
│ ├── metrics/
│ │ └── calculator.py # Behavioral metrics engine
│ ├── session/
│ │ └── manager.py # Session lifecycle
│ ├── privacy/
│ │ └── filter.py # Sensitive data redaction
│ └── config.py # Configuration
├── scripts/
│ ├── start_logger.bat # Windows startup
│ ├── manage_screenshots.py # Screenshot utility
│ └── view_metrics.py # Metrics dashboard
├── requirements.txt
└── README.md
git clone https://github.com/talha828/meero.git
cd meero
pip install -r requirements.txt| Platform | Command |
|---|---|
| Windows | Download from UB-Mannheim/tesseract |
| macOS | brew install tesseract |
| Linux | sudo apt install tesseract-ocr |
python -m logger.mainpython view_metrics.pyEdit logger/config.py to customize behavior:
| Setting | Description | Default |
|---|---|---|
ACTIVITY_POLL_INTERVAL |
Window check frequency | 2 sec |
SCREENSHOT_INTERVAL |
Screenshot capture frequency | 60 sec |
SCREENSHOT_POLICY |
keep_all, delete_after_ocr, keep_on_text |
delete_after_ocr |
IDLE_THRESHOLD_SECONDS |
Seconds before considered idle | 60 |
IGNORE_APPS |
Applications to exclude | ["Task Manager"] |
ENABLE_KEYBOARD_TRACKING |
Track typing speed | True |
| Table | Purpose |
|---|---|
sessions |
Start/end times, active/idle duration |
window_events |
Active window changes |
browser_events |
URLs and domains visited |
screenshots |
OCR text and file references |
clipboard_events |
Clipboard content (hashed) |
typing_metrics |
Keys per minute by application |
error_metrics |
Errors detected in window titles |
context_switches |
App switching frequency |
idle_recovery |
Recovery time after breaks |
tab_explosions |
Rapid tab switching events |
hourly_metrics |
Aggregated productivity scores |
- ✅ 100% Local – No telemetry, no cloud uploads
- ✅ Sensitive Data Redaction – Emails, passwords, credit cards auto-filtered
- ✅ Pause Anytime –
PAUSE_LOGGING = Truestops all tracking - ✅ App Ignore List – Exclude private applications
- ✅ Hashed Clipboard – Only SHA-256 stored, not raw content
- SQLite + WAL Mode – Concurrent database access
- Tesseract OCR – Screen text extraction
- OpenCV – Image preprocessing for OCR
- UI Automation – Cross-browser URL detection
- pynput – Keyboard monitoring
- Multi-threading – Non-blocking background tasks
- Multi-threaded activity tracking
- OCR-based browser URL extraction
- Screenshot + text retention policies
- Behavioral metrics (typing speed, errors, context switches)
- Privacy filters and local-only storage
- Web-based metrics dashboard (Streamlit)
- Weekly email reports
- System tray icon with controls
- Export to CSV/JSON
- Temporal pattern miner (frequent subsequence discovery)
- Causal graph builder (Granger causality on event streams)
- Latent state estimator (valence/arousal tracking)
- Complementary persona engine
- Multi-armed bandit for trigger optimization
Meero is Phase 1 of a larger ambition.
The current system is a comprehensive Raw Event Stream—it captures every window switch, keystroke rhythm, browser domain, idle pattern, and contextual screenshot. This is the foundational layer for something much more intelligent.
HPAPDA is a conceptual blueprint for second-generation AI companions—systems that don't just log what you do, but infer why you do it and adapt their personality in real-time to complement yours.
| Phase 1: Meero (This Repo) | Phase 2: HPAPDA |
|---|---|
| Raw event streaming | Temporal pattern mining |
| SQLite storage | Causal graph builder |
| Behavioral metrics | Latent state estimator |
| Rule-based nudges | Complementary persona engine |
- Pattern Discovery: "After 45 minutes of coding, user opens Twitter for exactly 3 minutes" → variable:
coding_fatigue - Causal Inference: Does checking Slack cause typing speed to drop? Or do both happen when you're already distracted?
- Dynamic Persona: An AI that becomes more stable when you're anxious, or gently initiates when you're withdrawn
- Local-First Intelligence: No cloud, no subscription, no privacy nightmares
📖 Read the full vision: HPAPDA: A Conceptual Blueprint for Second-Generation AI Companions
Meero is the necessary first step—the eyes and ears. HPAPDA is the brain that learns to be a decent, sarcastic, occasionally annoying but ultimately helpful second self.
Interested in contributing to Phase 2? Open an issue with the tag hpapda or reach out directly.
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
This project is licensed under the MIT License – see the LICENSE file for details.
This tool is intended for personal productivity tracking only. Do not use it to monitor others without their explicit consent. The author assumes no responsibility for misuse.
- Tesseract OCR – Open source OCR engine
- UI Automation – Windows accessibility framework
- All open source contributors who made this possible
Talha Iqbal – @Linkedin – talha.developer.01@gmail.com