Hear2Help

main page
example of what an emergency page looks like

Inspiration

During emergencies, communication can be the difference between life and death. Here's a scary thought: what happens during an emergency if you can't speak? Millions of people who are deaf, hard-of-hearing, non-verbal, or experiencing a panic attack are forced to rely on systems that assume verbal communication. Texting 911 isn’t widely available, and long messages are hard to write in moments of panic. We wanted to build something that lets anyone ask for help fast, clearly, and without needing to speak or type.

What it does

Hear2Help is an accessibility-first, minimal-touch emergency assistant designed for users who cannot communicate verbally. It works in just a few taps and provides visual, tactile, and AI-powered support. Key Features

Tap-Based Emergency Mode with large, color-coded buttons
AI Speech Generation (ElevenLabs) that speaks on behalf of the user
Location Awareness to show nearby hospitals/police stations
Accessibility-Optimized UI (high contrast, large tap areas, dyslexia-friendly fonts)
Calm Mode for panic attacks (breathing guide + grounding prompts)
Triple-Tap Emergency Shortcut to instantly trigger SOS mode
Quick Medical Profile for bystanders
Multilingual Interface

How we built it

Tech Stack: Typescript + Next.js Speech: ElevenLabs real-time TTS API ChatBot: Gemini API Translation: Google Translate Widget Location: Geolocation API + WiFi/GPS heuristics Design: WCAG-inspired color palette, accessibility testing with screen readers

Challenges we ran into

Designing a UI that works under high stress without overwhelming users
Getting real-time speech generation to be clear in noisy environments
Implementing the triple-tap gesture reliably
Coding-specific challenges such as getting translation to work, enabling geolocation services, etc
Making accessibility decisions without compromising simplicity

Accomplishments that we're proud of

A fully working emergency assistant that requires zero speaking or typing and can actually help people in real-life emergencies
Real-time AI speech that sounds natural and helps users communicate instantly
Successfully implemented the triple-tap SOS gesture
A real-time chatbot that can help answer any emergency questions
Translation services in any language
Geolocation services that show the nearest hospital, police station, landmark, and your own location

What we learned

Accessibility is not just a feature since it changes every design decision
Simplicity is surprisingly hard when designing for panic scenarios
Speech synthesis is more powerful than we expected for assistive tech
Building for diverse disabilities taught us to rethink assumptions we make on the daily