Speech Spark

An AI-Powered Pronounciation Coach for children to help improve their overall speech in a fun and interactive way. Children can read aloud short sentences and get real time feedback!

Comment

Inspiration

Speech Spark was inspired by the need to support people who struggle with speech confidence — from children learning to speak clearly, to individuals overcoming social anxiety, speech impediments, or learning English as a second language. We wanted to make speech practice feel safe, friendly, and fun.

What it does

Speech Spark is an AI-powered pronunciation coach that helps users improve their speech through interactive voice feedback.

Users can:

  • Select a friendly AI coach (powered by ElevenLabs)
  • Read aloud a short sentence and upload a voice clip
  • Receive personalized pronunciation feedback from Google Gemini
  • Hear encouraging spoken responses to help them improve
  • Speech Spark turns speech therapy into an engaging, game-like experience.

How we built it

We built Speech Spark using:

  • React Native (Expo) for the front-end user interface
  • Node.js + Express for the backend server
  • Google Gemini API for speech transcription and analysis
  • ElevenLabs API for text-to-speech voice generation
  • MongoDB for future scalability and user data management

Challenges we ran into

  • Setting up and authenticating multiple AI APIs
  • Managing media uploads and audio playback
  • Coordinating frontend-backend integration with environment variables
  • Maintaining project organization and version control

Accomplishments that we're proud of

  • Created a fully functioninig interactive demo.
  • Successfully connected the frontend and backend APIs
  • Designed a child-friendly user experience with animated feedback icons

What we learned

  • How to structure a full-stack AI app
  • Working with API pipelines
  • The importance of UX design for accessibility and engagement
  • How to make AI interactions feel personal, friendly, and safe

What's next for Speech Therapist Aid

  • Real-time voice recording and feedback
  • Training custom AI models on user speech patterns
  • User authentication and profile tracking
  • Connecting users with certified speech therapists
  • Expanding multilingual support for ESL learners
Share this project:

Updates