SignBridge

Inspiration

Communication barriers between spoken language and sign language limit access to information for Deaf and hard-of-hearing communities. We wanted to build a tool that makes everyday conversations and digital content more inclusive.

What it does

SignBridge converts spoken English into text, simplified language, SignWriting, and sign animation in real time. It works with microphone and system audio, making it useful for conversations, meetings, and media.

How we built it

We built SignBridge using Whisper for speech recognition, Groq-powered LLaMA models for language simplification, and AI-based sign translation. The app uses Edge AI to run locally with offline support across multiple platforms.

Challenges we ran into

Creating accurate speech-to-sign translation, optimizing AI models for offline performance, and supporting different devices while maintaining speed and usability were major challenges.

Accomplishments that we're proud of

We built a cross-platform accessibility solution that brings real-time spoken content conversion closer to users while prioritizing privacy through offline AI processing.

What we learned

We learned that accessibility-focused AI requires not only technical accuracy but also thoughtful design around real-world communication needs and user experience.

What's next for SignBridge

We plan to improve sign translation accuracy, expand language support, enhance sign animations, and bring more real-time accessibility features to everyday communication.

Built With

• groq
• llama
• whisper