AI-based Meeting Notes Generator using Whisper and BERT

In a workplace environment where documenting meetings has become a requirement to complete work and track decisions, we present the intelligent system, Audio's AI Meeting Notes Generator, which enables users to automatically transcript and summarize recorded meetings, capturing important discussion points, conclusions and action points, as raw audio data requiring processing by transformer-based models. The audio notes generation combines OpenAI's Whisper API to provide speech-to-text transcription accuracy with BERT for the natural language decoder to unfold natural language. An interactive web-based front end allowing the user to upload recorded meetings and see their meeting minutes organized in real time is built around the main component React.js. The backend, which can be built in Python using Flask or FastAPI, can focus on integrating AI and managing workflows, and interfacing with databases (PostgreSQL or MongoDB) handling the end-state data. This system always "Win's" determination no lost important data, plus provides a productive time-saver, while significantly reducing manual labour.