FocusLens

FocusLens is a macOS desktop application that improves eye contact and engagement during video calls. It runs fully locally with no cloud processing, ensuring complete user privacy while providing real-time gaze tracking, active speaker display, smart nudging, and post-call analytics.

Stars
6
Language
TypeScript
License
MIT
Updated
View on GitHub

FocusLens

Privacy-First Real-Time Eye Contact Assistant

FocusLens is a macOS desktop application that improves eye contact and engagement during video calls. It runs fully locally with no cloud processing, ensuring complete user privacy while providing real-time gaze tracking, active speaker display, smart nudging, and post-call analytics.

Features

Real-Time Gaze Tracking

  • Eye contact detection with <15 degree threshold
  • Head pose estimation (yaw, pitch, roll)
  • Pupil and eye center tracking
  • ~60 FPS inference with <30ms latency

Dynamic Overlay System

  • Compact Mode: Minimal floating bubble near webcam
  • Expanded Mode: Full metrics and controls
  • Speaker-Focus Mode: Active speaker prominently displayed
  • Analytics Mode: Real-time engagement visualization

Smart Nudging

  • Subtle visual cues when gaze drifts
  • Directional glow effects
  • Non-interruptive guidance during calls

Post-Call Analytics

  • Attention timeline visualization
  • Gaze heatmaps
  • Engagement graphs
  • Personalized improvement suggestions

Meeting Detection

  • Auto-detects Zoom, Google Meet, Microsoft Teams, Discord, Slack, FaceTime, WebEx, BlueJeans
  • Automatic session recording
  • Platform-specific metadata

Calibration System

  • 9-point grid calibration
  • Polynomial regression mapping
  • Cross-validation for accuracy

Privacy Controls

  • Pause/resume tracking (global hotkey)
  • Export and delete telemetry
  • Webcam permission management
  • 100% local processing

Architecture

Tech Stack

LayerTechnology
Desktop RuntimeTauri 2.x
BackendRust
FrontendNext.js 14 + React 18 + TypeScript
StylingTailwind CSS
AnimationsFramer Motion
DatabaseSQLite
GPU AccelerationMetal (Apple Silicon)

Project Structure

FocusLens/
├── apps/
│   └── desktop/              # Tauri desktop application
├── packages/                # Shared packages (future)
├── native/
│   └── rust-core/          # Rust backend modules
├── models/                  # ML model files
├── SPEC.md                  # Detailed specification
└── README.md

Installation

Prerequisites

  • macOS 13.0 (Ventura) or later
  • Rust 1.70+
  • Node.js 18+

Build from Source

# Clone the repository
git clone https://github.com/Manavarya09/FocusLens.git
cd FocusLens

# Install dependencies
npm install

# Build the frontend
cd apps/desktop
npm run build

# Build the Tauri application
cd src-tauri
cargo tauri build

Running in Development

cd apps/desktop
npm run dev

Or use the Tauri CLI:

cd apps/desktop/src-tauri
cargo tauri dev

Usage

  1. Launch FocusLens - The app opens to the main dashboard
  2. Grant Permissions - Allow camera and screen recording access
  3. Start Tracking - Click "Start Tracking" to begin gaze detection
  4. Join a Meeting - FocusLens auto-detects meeting platforms
  5. View Analytics - Check the Analytics tab after meetings

Keyboard Shortcuts

  • Cmd+Shift+P - Pause/Resume tracking
  • Cmd+Shift+O - Toggle overlay visibility

Configuration

Settings are stored in ~/Library/Application Support/FocusLens/settings.json:

{
  "overlay_mode": "compact",
  "overlay_position": { "x": 0.5, "y": 0.05, "monitor": 0 },
  "eye_contact_threshold": 15.0,
  "nudge_enabled": true,
  "auto_start_meetings": true,
  "target_fps": 60
}

Data Storage

All telemetry is stored locally in:

  • Sessions: ~/Library/Application Support/FocusLens/telemetry.db
  • Logs: ~/Library/Logs/FocusLens/
  • Settings: ~/Library/Application Support/FocusLens/settings.json

Security

  • No cloud processing - all inference runs locally
  • No data leaves your machine
  • Telemetry can be exported or deleted at any time
  • Camera access is only used for gaze tracking

Contributing

Contributions are welcome! Please read our contributing guidelines before submitting PRs.

Development Setup

# Install Rust
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# Install Node.js (using nvm)
nvm install 18
nvm use 18

# Install dependencies
npm install

# Run tests
npm test

Code Style

  • Rust: Follow rustfmt conventions
  • TypeScript: Follow ESLint rules
  • Commit messages: Conventional Commits format

License

MIT License - see LICENSE for details.

Acknowledgments

Support