Transcript Cleaner

Paste a raw meeting or video transcript and get clean, readable text. Strips timestamps in common formats, optionally removes speaker labels and filler words, and collapses stutter repeats, ideal before feeding a transcript to an LLM for summarizing.

Raw transcript (before)

Cleaned (after)

Strip timestamps ([00:12:34], 00:12, (12:34), SRT) Remove speaker labels (Name:) Remove filler words (um, uh, you know...) Collapse repeated words (I I I) Collapse extra blank lines

Timestamps removed

Fillers removed

Speaker labels removed

Words after cleaning

Characters removed

What gets cleaned: timestamps like [00:12:34], (12:34), bare 00:12 at the start of a line, and full SRT/VTT cue lines (index numbers and 00:00:01,000 --> 00:00:04,000 arrows). Speaker labels are lines starting with a short name followed by a colon, such as "Sarah:" or "Speaker 2:"; keep them on if you need to know who said what. Filler removal targets um, uh, er, ah, hmm, "you know", "I mean", "sort of", "kind of", and comma-wrapped "like". Review the output before publishing: filler removal is aggressive by design and can occasionally clip a legitimate use of "like".