Inspiration
From my family
What it does
MediVoice is an AI accessibility app that uses Google Gemini and ElevenLabs to scan medicine bottles, simplify the medical jargon, and speak the instructions out loud for elderly and visually impaired patients.
How we built it
React/Vite + Tailwind frontend, FastAPI + SQLAlchemy/SQLite backend, Gemini vision, ElevenLabs .
Challenges we ran into
- Windows file-path limits broke the ElevenLabs SDK setup.
- It took heavy tuning to get the AI to analyze multiple label photos, translate the text to Hindi, and keep the drug names in Latin.
Accomplishments that we're proud of
A true end-to-end flow: snap one or more label photos → Gemini reads them together → clear, large-print explanation with severity-coded warnings and expiry badges → one tap to hear it in English or Hindi. We're especially proud that it serves elderly and visually-impaired users..
What we learned
How to get reliable structured output from Gemini and ElevenLabs. We also learned a lot about Windows dev environment.
What's next for MediVoice
We will update it further for more easy use and more future improvements.
Log in or sign up for Devpost to join the conversation.