SONARI

Inspiration

SONARI was inspired by the challenge students and professionals face in retaining spoken information from lectures, meetings, and training sessions. Too often, valuable insights are lost simply because no one is able to take perfect notes. We envisioned a solution that listens, learns, and teaches back — turning every voice into structured, accessible, and actionable knowledge.

What it does

SONARI transforms any audio recording into rich, interactive study materials. With one upload or recording, it automatically:

Transcribes speech using AI-powered Whisper.
Summarizes content into concise notes and key points.
Generates flashcards, quizzes, and mind maps for active recall.
Offers an AI Tutor that explains difficult topics.
Provides a text-to-speech accessibility mode for the hard-of-hearing. SONARI is designed for both students and professionals, whether in classrooms, boardrooms, or workshops.

How we built it

SONARI is built with a React (Vite) frontend and a Flask (Python) backend. The backend uses the OpenAI API for transcription (Whisper) and summarization. Data is processed and structured into multiple output layers, summaries, quizzes, and flashcards. The frontend uses React Router for navigation across modules like Transcribe, Summarize, Study Tools, and AI Tutor. We integrated text-to-speech (TTS) and download options for user accessibility and convenience.

Challenges we ran into

Handling and validating multiple audio formats (.mp3, .wav, .m4a, etc.)
Parsing large AI JSON responses without breaking the frontend
Syncing transcription and summarization pipelines to reduce latency
Designing a UI that’s both intuitive and inclusive
Managing API limits and environment variables securely for team collaboration

Accomplishments that we're proud of

Successfully built an end-to-end AI learning assistant in under hackathon time constraints.
Integrated accessibility features that make learning inclusive.
Created a smooth multi-page React app with routing and dynamic components.
Developed a prototype that bridges education, accessibility, and productivity.

What we learned

How to deploy OpenAI’s transcription and summarization APIs efficiently.
How to architect and optimize a React + Flask project for real-time processing.
How to handle AI-generated JSON outputs and dynamically render them.
Most importantly, we learned that inclusive design is not optional, it’s essential.