Transcript Cleaner
Paste an auto-generated transcript and clean it up in one click. Remove filler words (um, uh, like, you know), fix sentence capitalization, add paragraph breaks every N sentences, and strip speaker labels and timestamps. See exactly how much was removed before you copy or download the cleaned text.
How to use
- 1
Paste your transcript
Drop in raw text from YouTube, Otter.ai, Descript or any auto-transcription service. Speaker labels, timestamps and filler words are all fine.
- 2
Toggle cleanup options
Pick which transformations to apply — filler removal, capitalization, paragraph breaks, speaker label removal, and timestamp removal. All are on by default.
- 3
Adjust paragraph length
When paragraph breaks are on, use the slider to choose how many sentences go in each paragraph (1–10). 3 is a good default for readability.
- 4
Review the stats and copy
See how many fillers, words, and bytes were removed. Copy the cleaned transcript or save it as a .txt file.
Common use cases
- Turn a raw podcast transcript into a clean blog post draft.
- Prepare a YouTube auto-transcript for use as an article.
- Strip filler words before feeding a transcript to an LLM for summarising.
- Clean up speaker-labelled interview transcripts for publication.
Limitations
- Filler-word removal is based on a fixed list of ~20 English phrases. It may miss uncommon fillers (e.g. "right" as a filler) and could over-strip the word "like" in legitimate uses.
- Capitalization is sentence-based — it capitalises the first letter after a period, exclamation or question mark. It does not fix proper nouns or correct ALL-CAPS shouting.
- Speaker-label removal uses a regex that matches common patterns like "Name:", "[Name]:" or ">> Name:" at the start of a line. Unusual formats may slip through.
- Timestamp removal strips bracketed timestamps and bare
MM:SStokens. It does not touch ISO-format dates or durations inside running prose. - No punctuation is added — the tool only cleans existing text. For auto-punctuation, use an LLM-powered transcription service.