I used this CLI tool for it: https://github.com/m1guelpf/auto-subtitle
Whisper-based, worked great for subtitling S2 of ReGenesis as for some reason the DVDs lacked subtitles. I used the largest model, I think. It even managed to correctly identify which character was crying, which I thought was impressive.
That depends? What exactly are your costs? My server at home costs pennies even with my crazy energy prices here. My Nextcloud, RSS, and random stuff VPS is only 15€ per quarter, Lemmy VPS is less than 5 per month, and Backblaze for backup is pennies per month. For me the only sensible cost-cutting would be to unite the two VPS to one beefier VPS. But that probably doesn’t make sense for you.