The most common problem for anyone singing outside the top 100 — and it is solvable, with caveats.
Karaoke catalogues are built on commercial demand. That predictably leaves out:
1. Ask. Karaoke communities have people who make tracks on request for free. It's slow and depends on a stranger's goodwill, but it costs nothing and the results are usually good.
2. Do it yourself. Separate the vocals with UVR and time the lyrics. Costs time — about an hour per song — and no money.
3. Automate it. A system that separates and transcribes on its own. Saves time, loses accuracy on the lyrics, and you end up editing.
If the only recording that exists is live, with an audible crowd and heavy reverb, the instrumental will have audible artefacts and the lyric transcription will get a lot wrong. It's doable, but go in with the right expectation: you will be fixing a lot by hand.
A distinction that confuses a lot of people: separating the vocals and having timed lyrics are two different jobs. The first is automatic now. The second, if you want the syllable highlighting on the beat, is still the hard part.
Three routes: ask in a karaoke community (free and slow), do it yourself by separating the vocals with UVR and timing the lyrics (about an hour per song), or use a system that does both automatically, accepting that the lyrics will need correcting.
You can, but it is the worst case. Crowd noise and room reverb hurt both the vocal separation and the lyric transcription. The instrumental comes out with audible artefacts.
Rarely. Cast recordings almost never have karaoke versions because the market is too small to justify one. The route there is separating the vocals from the cast album itself — and the clear diction of that kind of recording helps a lot.
StarSinger does those three steps for you. You upload a file you already have, it separates the vocals, transcribes the lyrics with per-syllable timing and gives you a karaoke that plays in the browser, with pitch scoring. Free right now.
The caveats, because they're real: the transcription gets words wrong on screamed or heavily distorted vocals, and processing takes well over the length of the song.
Try it free