Accent Bridge

• 

  Accent Bridge Web App

• 

  Architecture Diagram

Inspiration

Accents often create unintended barriers in communication. Many people speak clearly, but listeners struggle due to accent differences rather than content. Accent Bridge was inspired by the idea that people should not have to change who they are to be understood.

What it does

Accent Bridge allows users to record their voice once and receive a clarity-optimized version of their speech in a selected accent (Neutral, US, or UK), while preserving the original meaning and intent.

It focuses on accent adaptation, not translation or voice alteration.

How we built it

The project uses: • Speech-to-Text to transcribe recorded audio • Text-to-Speech to re-synthesize the speech using accent-specific voices • A clean, demo-focused UI built with modern web technologies • Google Cloud Run for scalable, containerized deployment

Audio is processed ephemerally and never stored.

Challenges we ran into

• Differentiating accents clearly without changing tone or intent
• Handling short vs long recordings reliably
• Designing a UI that feels intuitive, fast, and demo-friendly
• Managing cloud authentication and deployment within hackathon time limits

Accomplishments that we're proud of

• A working end-to-end accent adaptation pipeline
• Clear accent selection with instant re-conversion
• A polished, privacy-first demo experience
• Successful cloud deployment using containerized infrastructure

What we learned

We learned how subtle accent differences can significantly impact understanding, and how careful orchestration of speech APIs can improve clarity without altering meaning.

What's next for Accent Bridge

• Live (real-time) accent adaptation
• Voice-preserving synthesis (closer to the original speaker)

Built With

• fastapi
• google-cloud
• next.js
• python
• tailwindcss
• typescript