Inspiration
Our project is an ==Auto-Karaoke== app, to quickly and easily get the karaoke track of the original song from YouTube. The Karaoke has a wide range of applications, from singing training over the track, to listening to music while you’re studying without voice interruptions.
What it does
The Karaoke app turns any YouTube link into a karaoke track, separating the vocals form the background, playing only the instrumental sound and the lyrics of the video.
How we built it
We created a front-end to drag and drop the URL from YouTube. Once we have the link from the user, the app download it and separate audio from video with YouTube-DLP. Next, it applied AI tool (MDX-Net, an audio separation model) to isolate the instrumental track. In parallel process, using genius.api tool, the App identify the lyrics from the original song. Finally, the interface capture and display the instrumental track and the lyrics.
Challenges we ran into
Connecting the front end and the back end, dealing with CORS, running all servers in all computers, and installing all packages required.
Accomplishments that we're proud of
It ended up actually working in the very end
What we learned
FASTapi, coordinating with teammates, the pain of linking everyone's code together so it works well.
What's next for Karaoke App
Timesynced lyrics, more robust lyric matching, improved UI/UX
Built With
- javascript
- python
- react
- tailwind
Log in or sign up for Devpost to join the conversation.