device-pair

🧠 CAPE: Context-Aware Policy Engine

Transform your Android device into an intelligent personal assistant that adapts to your life in real-time.

CAPE is a cutting-edge Android-first automation system that leverages AI to understand your context and automatically optimize your device behavior. By sensing calendar events, location, usage patterns, and biometric signals, CAPE coordinates with OpenClaw AI agents to apply the right "behavior pack" at exactly the right time.

🌟 Key Features

📱 Intelligent Context Sensing

Calendar Intelligence: Automatically reads upcoming meetings and prepares focus mode
Location Awareness: Detects home, office, college, and commuting patterns
Usage Analytics: Monitors screen time, app switching, and notification patterns
Sleep Tracking: Integrates sleep debt data into decision-making
Commute Pressure: Real-time traffic and travel time analysis

🤖 AI-Powered Decision Making

OpenClaw Integration: Leverages advanced AI agents for intelligent decisions
Stress Scoring: Real-time assessment of user cognitive load
Adaptive Learning: Improves recommendations based on user feedback
Multi-Agent Orchestration: Coordinates specialized AI agents for different contexts

🎯 Smart Automation Packs

Focus Mode: Blocks distractions during work/study sessions
Relax Mode: Optimizes device for downtime and recovery
Commute Mode: Provides navigation assistance and travel updates
Home Mode: Creates a comfortable personal environment

🔒 Privacy-First Architecture

Local Processing: All data processed locally on your machine
USB Bridge: Secure communication between phone and gateway
No Cloud Dependencies: Your data never leaves your devices
Open Source: Fully transparent and auditable codebase

🚀 Quick Start

Prerequisites

Node.js 22+ (for gateway server)
JDK 25+ (required for Android build)
Android SDK with ADB (for phone communication)
Android device with USB debugging enabled
Google Maps API Key (for location services)

Installation & Setup

🍎 macOS / Linux

# Clone and install dependencies
git clone <repository-url>
cd CAPE2
npm install

# Build the Android APK
./tools/build-apk.sh

# Start the gateway server (Terminal 1)
./tools/start-gateway.sh

# Install and launch app (Terminal 2)
./tools/install-apk.sh

🪟 Windows

# Install JDK 25 (if not already installed)
winget install Microsoft.OpenJDK.25

# Clone and install dependencies
git clone <repository-url>
cd CAPE2
npm install

# Build the Android APK
.\tools\build-apk.ps1

# Start the gateway server (Terminal 1)
.\tools\start-gateway.ps1

# Install and launch app (Terminal 2)
.\tools\install-apk.ps1

🔧 Configuration

1. Enable USB Debugging on your Android device:

Settings → About Phone → Tap "Build Number" 7 times
Settings → Developer Options → Enable "USB Debugging"

2. Connect your device via USB and authorize debugging

3. Configure Environment (required):

   cp .env.example .env
   # Edit .env with your preferences and API keys

Required Environment Variables:

   GOOGLE_MAPS_API_KEY=your_google_maps_api_key_here
   OPENCLAW_BASE_URL=http://127.0.0.1:18789  # Optional
   OPENCLAW_TOKEN=your_openclaw_token_here        # Optional

📊 Architecture Overview

┌─────────────────┐    USB/ADB     ┌─────────────────┐    HTTP     ┌─────────────────┐
│   Android App   │ ◄────────────► │  Gateway Server │ ◄──────────► │  OpenClaw AI    │
│                 │                │   (Node.js)     │             │   Agents        │
│ • Context Sense │                │                 │             │                 │
│ • UI Interface  │                │ • Decision API  │             │ • Stress Scoring│
│ • Automation    │                │ • Feedback Loop │             │ • Learning      │
└─────────────────┘                │ • Memory Store  │             │ • Orchestration │
                                   └─────────────────┘             └─────────────────┘

Core Components

packages/cape-core: JavaScript reference implementation for stress scoring and decision logic
packages/cape-gateway: HTTP server handling communication between Android and AI agents
android/: Kotlin/Jetpack Compose application with modern UI
openclaw/: AI agent configurations, memory stores, and skill definitions

🎮 Usage Guide

First-Time Setup

1. Launch the CAPE app on your Android device 2. Complete onboarding with your name and frequently visited places 3. Grant permissions for location, calendar, notifications, and usage access 4. Choose your current context (Home, Office, College, Commute, or Relaxing) 5. Apply your first behavior pack and experience intelligent automation

Daily Use

Automatic Detection: CAPE automatically detects when you arrive at saved locations
Smart Transitions: Device behavior adapts as you move between contexts
Feedback Loop: Provide feedback to improve AI recommendations
Commute Assistance: Get real-time traffic updates and route suggestions
USB Sync: Automatic connection maintenance via keep-phone-synced.ps1
Local Processing: All data stays on your devices, no cloud dependencies

🧪 Testing

Core Logic Tests

npm test

Gateway Tests

npm run gateway:test

Demo Scenarios

npm run demo:scenarios

🔌 API Endpoints

The gateway server exposes RESTful APIs at http://127.0.0.1:8787:

Core Endpoints

POST /v1/context/decision - Get AI-powered context decisions
POST /v1/feedback - Submit user feedback for learning
GET /v1/openclaw/status - Check AI agent status

Location Services

POST /v1/maps/geocode - Convert addresses to coordinates
POST /v1/maps/search - Search for places by name

🤖 OpenClaw Integration

CAPE leverages the OpenClaw personal AI assistant framework for advanced decision-making:

Setup Options

Option 1: Full OpenClaw Integration (Recommended)

# Install OpenClaw
npm install -g openclaw@latest
openclaw onboard --install-daemon
openclaw gateway --port 18789 --verbose

# Bootstrap CAPE workspace
node tools/openclaw-bootstrap.mjs

# Start CAPE with OpenClaw brain
OPENCLAW_REQUIRED=true OPENCLAW_BASE_URL="http://127.0.0.1:18789" OPENCLAW_TOKEN="your-token" npm run gateway:start

Option 2: Local Fallback Mode

# Uses built-in decision logic when OpenClaw is unavailable
npm run gateway:start

Configuration

Configure OpenClaw integration via environment variables:

OPENCLAW_REQUIRED=true - Require OpenClaw for all decisions
OPENCLAW_BASE_URL - OpenClaw gateway URL
OPENCLAW_TOKEN - Authentication token
OPENCLAW_AGENT_ID - Agent identifier (default: "cape")

� Direct APK Installation

For quick testing without building from source:

1. Download Cape.apk from repository 2. Enable USB debugging on Android device:

Settings → About Phone → Tap "Build Number" 7 times
Settings → Developer Options → Enable "USB Debugging"

3. Install APK: adb install Cape.apk 4. Start gateway server: .\tools\start-gateway.ps1 5. Connect phone via USB - app will automatically sync

�� Project Structure

CAPE/
├── CAPE2/                   # Main development workspace
│   ├── packages/
│   │   ├── cape-core/          # Core decision logic and stress scoring
│   │   └── cape-gateway/       # HTTP gateway server
│   ├── android/                # Android application source
│   ├── openclaw/               # AI agent configurations
│   ├── tools/                  # Build and deployment scripts
│   └── .env.example            # Environment configuration template
├── Cape.apk                 # Pre-built APK for distribution
├── README.md                 # This file
└── openclaw/                 # OpenClaw workspace and memory

🔧 Development

Building from Source

# Install dependencies
npm install

# Build all packages
npm run build

# Run in development mode
npm run dev

USB Connection Maintenance

The system includes automatic USB connection management:

keep-phone-synced.ps1 maintains gateway and ADB reverse port forwarding
Runs in background to ensure continuous phone-gateway communication
Automatically restarts gateway if connection is lost
Monitors device connection status every 10 seconds

Contributing

1. Fork the repository 2. Create a feature branch 3. Implement your changes with tests 4. Submit a pull request with detailed description

🎯 Use Cases

For Professionals

Meeting Preparation: Automatically enables focus mode before important calls
Work-Life Balance: Smoothly transitions between work and home environments
Productivity Optimization: Reduces distractions during deep work sessions

For Students

Study Sessions: Creates optimal learning environment during class/homework time
Campus Navigation: Assists with commute between campus locations
Exam Preparation: Minimizes distractions during critical study periods

For Everyone

Digital Wellness: Promotes healthy device usage patterns
Sleep Quality: Reduces evening blue light exposure
Stress Management: Adapts device behavior based on cognitive load

🏆 Awards & Recognition

Built for Samsung PRISM OpenClaw Hackathon - Demonstrating the future of personal AI assistants that truly understand and adapt to human behavior.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Support

Documentation: See /docs directory for detailed architecture notes
Issues: Report bugs and request features via GitHub Issues
Community: Join our Discord community for discussions and support

---

CAPE: Where AI meets everyday life. Experience the future of personal automation today. 🚀

Summary

Install to Claude Code