Clean up a Zoom or Teams transcript
Paste a .vtt or .srt transcript, or drop the file — get clean speaker paragraphs and a compact version ready to paste into ChatGPT, with a live token count so you can see how much you just saved.
No transcript yet. Paste one above, drop a file, or load a realistic example to see timestamps stripped, speakers merged, and the token count drop.
How this works & why it's private
Everything happens with plain text parsing, right here in this tab. When you paste or drop a transcript, we split it into caption "cues" using regular expressions — the same kind of pattern-matching a spreadsheet's find-and-replace uses — and never send the text anywhere. Timestamps, speaker tags like <v Speaker Name>, and cue numbers are stripped locally; consecutive lines from the same speaker are joined into paragraphs; and repeated caption fragments (the rolling, half-finished lines Zoom and Teams save while a sentence is still being spoken) are collapsed by comparing word overlap between consecutive cues.
The token count is a simple estimate — characters divided by four, a common rule of thumb — not a call to an official tokenizer, because that would mean sending your text to a server. We say "estimate" because it is one.
To verify none of this touches a network: open your browser's dev-tools Network tab before you start, paste a transcript, and watch the live "uploads" counter in the nav above — it will stay at 0.