dimatorzok.com

How to Remove “Subtitles by DimaTorzok”

Fastest way — paste the text or upload the subtitle file into the cleaner: it finds known hallucination phrases and removes them while keeping timecodes. Everything runs in your browser; files are never uploaded to a server.

Telegram: voice-message transcript

If the phrase appeared in a transcribed voice message, it's a recognition glitch on silence at the start/end of the recording. Telegram doesn't let you edit the transcript, so:

  1. copy the transcript text;
  2. paste it into the cleaner and click “Find hallucinations”;
  3. copy the cleaned text.

CapCut / Instagram Edits / DaVinci: auto-captions

In video editors the credit lands in the caption track as a separate segment over a pause. Options:

  1. find the caption segment over the pause and delete it right in the editor;
  2. or export captions to .srt, run them through the cleaner and re-import.

An .srt or .vtt file

  1. upload the file to the cleaner (or drag it into the field);
  2. review the flagged lines — uncheck anything you want to keep;
  3. click “Download clean file” — numbering and timecodes are recalculated automatically.

Your own Whisper pipeline

If you transcribe yourself and want hallucinations gone at the source:

Need recognition set up for a business (bots, subtitles, media monitoring) without these errors? Get in touch — see contacts.

FAQ

Will the cleaner delete my real text?

Not if you review the list. It works in a “show and let you decide” mode: it only highlights suspicious lines and removes only the ones still checked. Unique credits (DimaTorzok, Amara.org, etc.) are matched precisely; generic phrases like “thank you for watching” are removed only when they make up the entire line.

What if my phrase is in another language?

The cleaner catches signatures in several languages. If yours is missing, add it in the “your phrase” field and the tool will remove it too.