If you make voiceover videos, you already know the quiet problem: modern AI tools don't just read your text — they touch it. A model tightening a script, a text-to-speech engine reshaping a line, a translation step dropping a clause it decided you didn't need. Usually it's small. Sometimes it changes the meaning. And it's rarely consistent twice in a row. When you're producing at volume, you can't stand over every line to check.

We built SRTisan because we were living inside that problem — cleaning up ElevenLabs voiceover output and wrestling SRT subtitle files by hand, one faceless video at a time. So we made three small, careful tools and put them in one place. The first splits your translated text into clean, well-sized lines before it ever reaches your voiceover engine. The second repairs the damage ElevenLabs does when it breaks your carefully chosen line splits to fit its own character limits — restoring your intended lines with their real timestamps intact. The third cleans the .srt that comes out at the end — stray periods, orphaned words, broken timing — so your subtitles read like a person made them.

The important part: there's no AI in the critical path. SRTisan does the boring, exact work with plain, deterministic rules and checks every character it touches, so what goes in comes back out. That reliability is what lets it scale subtitle localization and faceless video production without adding risk — in our own pipeline it cut manual cleanup enough to lift throughput by up to 60%.

Three small tools, doing one honest thing: protecting the words you already wrote.