AccessiComm is a comprehensive mobile Android application designed to facilitate seamless communication between deaf, blind, and speech-impaired individuals, as well as with the general public. The app leverages machine learning, computer vision, and multimodal interaction to break down communication barriers.
To create an inclusive communication platform serving 466M+ deaf individuals and 285M+ visually impaired individuals globally, establishing AccessiComm as the leading accessibility communication tool.
- Real-time sign language recognition using advanced ML models
- Text/Speech to Sign Language animation with 3D avatars
- Interactive learning mode with accuracy feedback
- Support for multiple sign languages (ASL, ISL, BSL)
- Object & Scene Recognition with spatial audio feedback
- OCR Text Reading for documents, menus, and signs
- Face Recognition for familiar people
- AI-powered scene description with natural language
- Text, Voice, Sign Video, and Image messaging
- Real-time video calls with live translation
- Quick communication mode with preset phrases
- Group communication with mixed accessibility needs
- SOS Emergency Alert with location sharing
- Medical ID & Information display
- Emergency contact management
- Quick emergency phrases
- WCAG 2.1 Level AA compliance
- High contrast mode and large text support
- Screen reader compatibility (TalkBack)
- Voice commands for all features
- Customizable haptic feedback
- Offline mode for critical situations
- UI Framework: Jetpack Compose with Material 3
- Architecture: MVVM with Clean Architecture
- Dependency Injection: Hilt
- Database: Room (SQLite)
- Networking: Retrofit + OkHttp
- ML/AI: TensorFlow Lite, MediaPipe
- Backend: Firebase (Auth, Firestore, Storage, Messaging)
- Navigation: Navigation Compose
app/
├── src/main/java/com/accessicomm/app/
│ ├── data/ # Data layer
│ │ ├── local/ # Room database
│ │ ├── model/ # Data models
│ │ └── repository/ # Repository implementations
│ ├── di/ # Dependency injection
│ ├── ui/ # UI layer
│ │ ├── components/ # Reusable UI components
│ │ ├── navigation/ # Navigation setup
│ │ ├── screens/ # Screen composables
│ │ └── theme/ # App theming
│ └── service/ # Background services
- Android Studio Hedgehog or later
- Android SDK 26+ (Android 8.0+)
- Kotlin 2.0.21+
- Gradle 8.13.0+
- Clone the repository
- Open in Android Studio
- Sync project with Gradle files
- Run on device or emulator
- Add Firebase configuration (
google-services.json) - Configure ML model files in
assets/ml_models/ - Set up API keys in
local.properties
- Uses ASL for communication
- Needs real-time translation for hearing individuals
- Requires visual alerts for sounds
- Uses screen readers and assistive technology
- Needs object detection and scene description
- Requires OCR for reading documents
- Uses text-to-speech communication
- Needs quick phrase libraries
- Requires emotion expression in messages
- Learning sign language basics
- Facilitating communication for family members
- Managing emergency contacts and medical info
- Accessibility First: Every screen usable by all disability types
- Minimal Complexity: Maximum 3 taps to reach any feature
- Clear Feedback: Visual + Audio + Haptic for all actions
- Consistency: Unified design language across all screens
- Forgiveness: Easy undo, clear confirmations for destructive actions
./gradlew assembleDebug./gradlew test
./gradlew connectedAndroidTest./gradlew lintDebug- App launch: <3 seconds (cold start)
- Feature activation: <1 second
- ML inference: <500ms per frame
- Message delivery: <2 seconds
- RAM usage: <400MB during active use
- Battery drain: <20% per hour of active use
- End-to-end encryption for messages
- Local processing for camera/ML (no image upload by default)
- Encrypted storage for sensitive data
- GDPR and CCPA compliant
- Granular permissions with clear explanations
- Phase 1: English, Spanish, Hindi, Mandarin
- Phase 2: +10 additional languages
- Sign languages: ASL, ISL, BSL (phased rollout)
- 50,000+ downloads within 6 months
- 4.5+ star rating on Google Play Store
- 60%+ user retention at 30 days
- Featured in Google Play's accessibility collection
- Partnership with 3+ major accessibility organizations
We welcome contributions! Please see our Contributing Guidelines for details.
This project is licensed under the MIT License - see the LICENSE file for details.
- Email: support@accessicomm.app
- Website: https://accessicomm.app
- Documentation: https://docs.accessicomm.app
- Google's accessibility team for guidance
- The deaf and blind communities for feedback
- Open source ML model contributors
- Accessibility advocates and organizations
AccessiComm - Breaking down communication barriers with AI-powered accessibility.