Bridge

Bridge — giving every voice a way to be heard.

Comment

Inspiration

Millions of people with speech disabilities face barriers in everyday communication — from ordering coffee to making a simple phone call. We wanted to build a bridge between voice and understanding — an AI-powered companion that empowers every individual to express themselves freely, no matter how they communicate.

Bridge began with a simple idea: technology should listen, speak, and empathize just like we do.

What it does

Bridge is an accessibility-first communication app that enables natural two-way interaction between speech-disabled individuals and others — whether face-to-face or over a call.

  • Text-to-Speech (TTS): Converts typed or pre-set messages into expressive, human-like voices using Android’s TTS engine.
  • Speech-to-Text (STT): Transcribes incoming speech in real time using Google’s SpeechRecognizer API.
  • Call Mode: Facilitates live phone conversations by seamlessly syncing TTS and STT.
  • Floating Assistant Button: Accessible from any app for quick responses and message playback.
  • AI Suggestions: Gemini AI generates smart, context-aware replies and natural phrasing assistance.

Bridge makes communication not only possible — but effortless, inclusive, and dignified.

How we built it

  • Frontend (Android): Java + XML using Material Design 3, dark neon UI, and accessibility-first layouts.
  • Backend: Spring Boot REST API with MongoDB Atlas for storing conversations and user profiles.
  • Speech Pipeline:
    • Google SpeechRecognizer API for transcription (STT).
    • Android Text-to-Speech Engine for natural voice output.
  • AI Layer: Gemini-powered contextual suggestions and RAG-based personalization.
  • Threading: Multi-threaded architecture ensures smooth parallel processing of UI, STT, and TTS.
  • Analytics & Auth: Firebase for user authentication and session tracking.

Challenges we ran into

  • Achieving low-latency voice recognition on limited hardware.
  • Managing simultaneous STT, TTS, and UI rendering across threads.
  • Implementing a persistent floating assistant that works across apps and Android versions.
  • Ensuring network security (cleartext vs HTTPS) while maintaining performance.
  • Designing an interface that’s both minimal and intuitive for accessibility.

Accomplishments we’re proud of

  • Built a fully functional communication bridge between typed and spoken language.
  • Created a floating accessibility button that stays available across apps.
  • Integrated Gemini AI for smart, context-aware replies.
  • Delivered a clean, neon-themed Material UI optimized for readability and speed.
  • Designed a system architecture that scales — combining Android + Spring Boot + MongoDB seamlessly.

What we learned

  • Deep understanding of speech processing pipelines (STT, TTS, async synchronization).
  • How to design for accessibility-first UX — balancing simplicity with modern aesthetics.
  • Multi-threading and concurrency in Android for real-time operations.
  • Using AI to enhance communication, not replace it.

What’s next for Bridge

  • Emotion-aware AI voices that adapt tone and pace to emotion or context.
  • Multilingual and offline support for global accessibility.
  • Bridge Agent: a RAG-powered assistant using Gemini + MongoDB Atlas Vector Search for smarter, personalized replies.
  • Collaborate with assistive tech NGOs and release Bridge free for those in need.
  • Publish on Google Play Store after performance and accessibility audits.

Built With

Share this project:

Updates