VidGet is a full-stack video downloader web app powered by yt-dlp on a Python/FastAPI backend and a Next.js frontend. Users paste any video URL and instantly get a list of available quality options — from 4K to 360p, audio-only MP3, and multi-language audio tracks — which they can preview in-browser or download directly to their device.
Building a fast, reliable pipeline that streams video files from source to client without re-encoding, while correctly merging video-only and audio-only streams on the fly. The challenge also included handling hundreds of platforms through yt-dlp, exposing clean progress events to the frontend, and keeping the UI responsive during long downloads.
Delivered a production-ready downloader with a clean warm-toned UI, real-time download progress indicators, in-browser video preview via blob URLs, multi-language audio track selection, and a FastAPI backend that streams files directly — no re-encoding unless a video/audio merge is needed.
