AI-Powered Accessibility

Breaking barriers
with ARIA

A universal accessibility assistant that gives voice to the deaf and eyes to the blind — powered by real-time AI on the edge.

0 Accessibility Modes
0 AI Vision Providers
0 Languages Supported

Two modes. One mission.

ARIA adapts to your needs with two powerful accessibility modes, each designed for a specific community.

S
For Deaf & Mute

SIGN Mode

Sign in American Sign Language using your phone camera. ARIA recognizes your hand signs, detects your facial emotion, and speaks your words aloud with matching tone and feeling.

  • Real-time ASL recognition via AI vision
  • Facial emotion detection (8 emotions)
  • Emotion-matched text-to-speech output
  • Multi-language translation (EN, ES, FR, HI)
  • Live transcript with history
G
For Blind

GUIDE Mode

Point your camera ahead and ARIA warns you about obstacles in real-time with voice alerts and haptic feedback — keeping you safe as you navigate the world.

  • On-device YOLOv8 obstacle detection
  • 3-level severity (clear / caution / danger)
  • Instant voice warnings via TTS
  • Haptic feedback (heavy for danger)
  • Zero-latency local processing

Simple. Powerful. Instant.

ARIA processes everything in real-time on a Jetson Orin Nano — no cloud delay for safety-critical features.

1

Open your camera

Choose SIGN mode (front camera for your hands) or GUIDE mode (back camera for the path ahead). One tap to start.

2

AI processes in real-time

Frames stream to the Jetson backend at 5 FPS. AI vision providers recognize signs or detect obstacles in milliseconds.

3

Hear the result

SIGN mode speaks your translated sentence with emotion-matched voice. GUIDE mode alerts you with voice warnings and haptic vibration.

Built for the edge

Powered by cutting-edge AI models with intelligent fallback — running on NVIDIA Jetson for maximum speed and privacy.

AI

Multi-Provider Vision

Automatic fallback across Gemini, OpenAI, Claude, and local YOLOv8 for maximum reliability.

RT

Real-time Processing

WebSocket streaming at 5 FPS with on-device YOLO inference for zero-latency obstacle detection.

EM

Emotion Intelligence

FER + MTCNN detects 8 facial emotions. ElevenLabs speaks with matching tone and feeling.

HW

Edge Hardware

Runs on NVIDIA Jetson Orin Nano Super with CUDA 12.6 for GPU-accelerated inference.

ML

Dual ASL Pipeline

Rule-based MediaPipe classifier + ResNet50 neural network for robust sign recognition.

DB

Smart Analytics

Gemini-powered dashboard generates personalized insights from your usage patterns.

SOS

Emergency SOS

In a critical moment, tap the SOS button. ARIA instantly broadcasts an emergency message at full volume, records your GPS coordinates, and logs the event — giving you a lifeline when you need it most.

Get ARIA on your phone

We're preparing ARIA for the App Store and Google Play. Join the waitlist to be notified when it launches.

Coming soon on
App Store
Coming soon on
Google Play