BriefMe

What it does

BriefMe backend API has two versions, v1 and v2.
Version 1: Used Youtube DL to extract audio, Google Speech-to-Text for transcription, and a custom summarizer for summary generation. This process took about 3-4 minutes per video.
Version 2: Improved by extracting subtitles with Youtube DL and using Google Vertex AI for summarization. This reduced processing time to 5-10 seconds per video, improving performance by 97.3%.