Developer Vishal Mundewadi created AccessLearnAI, a pipeline designed to make lecture content accessible in multiple formats. The system takes a lecture recording and generates a text transcript using Whisper, a concise summary with T5, natural-sounding audio output via Edge TTS, and Braille formatting. The pipeline was built using Python with FastAPI for the backend, React and TypeScript for the frontend, and deployed on Google Cloud Platform with Vercel for hosting. A key challenge encountered was managing memory constraints on a free-tier deployment, where loading both Whisper and T5 models simultaneously exceeded available resources. AI
IMPACT Enables broader access to educational content by providing multiple output formats from a single lecture source.
RANK_REASON The item describes a specific application built using existing AI models and frameworks, rather than a new model release, research breakthrough, or significant industry event.
- AccessLearnAI
- Docker
- Edge TTS
- FastAPI
- Google Cloud Platform
- PostgreSQL
- React
- T5 Text To Text Transfer Transformer
- TypeScript
- Vercel
- Vishal Mundewadi
- Whisper
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →