- How do I make a karaoke video from any song?
- Upload the song and click Process. The pipeline separates the vocals from the instruments with AI source separation, runs speech recognition on the isolated vocal track to get every lyric with word-level timestamps, and renders a video where the lyric lines appear with each word highlighting in sync — over the instrumental. The whole thing is automatic.
- How accurate are the lyrics?
- Good on clear lead vocals — transcribing the ISOLATED vocal track (rather than the full mix) is what makes this work, since music normally defeats speech recognition. Heavily layered harmonies, screamed vocals, or dense rap can still produce errors; pop, rock, acoustic, and ballads transcribe best.
- Are the vocals completely removed?
- The instrumental comes from the same AI separation used by our vocal remover — on most studio recordings the vocal removal is clean, with at most a faint residue on heavily reverbed vocals. Backing vocals mixed like instruments may partially remain, which many karaoke singers actually prefer as a guide.
- What languages does it support?
- Lyric transcription covers 25 European languages including English, Spanish, German, French, Italian, and Portuguese. Songs in other languages will still get a clean instrumental, but the on-screen lyrics may be unreliable.
- Can I use my own lyrics file instead?
- Not yet — this version transcribes automatically. If the transcription misses lines in a complex song, a manual-lyrics (LRC/SRT) upload is the planned next step. For now, automatic transcription covers the typical karaoke-night use case.
- Is this legal for karaoke night?
- Making a karaoke version of a song you own for private use is generally fine. Public performance or re-uploading commercial songs as karaoke videos involves the song's copyright — venues typically carry performance licenses, and platforms may content-match the audio. For original or royalty-free music there's no issue.