Turn YouTube auto-captions or messy interview transcripts into clean, readable text in 3 clicks—no regex, no manual find-and-replace. This guide shows you how to remove timestamps, speaker labels, and filler words while preserving context, plus how to export clean SRT files with timing intact.
Table of Contents
Category hub: /creator/captions
Quick Start
- Open the Transcript Cleaner
- Paste your transcript or upload .txt/.srt
- Toggle ON: Remove timestamps, Remove speakers, Smart punctuation
- Enable "Remove filler words" and add custom fillers if needed
- Click "Clean it" and review the output
- Copy to clipboard or export (TXT always, SRT for SRT inputs)
What the Transcript Cleaner Fixes
Auto-captions and interview transcripts are rarely publish-ready. They come loaded with timestamps in multiple formats, speaker labels that break flow, filler words like "um" and "uh," and inconsistent punctuation. The Transcript Cleaner detects and removes these patterns automatically, so you can focus on editing content instead of hunting for timestamps.
Timestamps (All Formats)
The tool removes standard formats (HH:MM, HH:MM:SS), inline brackets ([00:12.34]), SRT time-range and index lines, YouTube-style (1:23), casual formats (1m23s, 1h2m), and "at 1:23" patterns. Toggle the option and every timestamp disappears in one pass.
Speaker Labels
Interview tools like Rev and Descript prefix lines with "Speaker 1:" or "John:" labels. The cleaner strips these labels while preserving the actual dialogue, so your transcript reads as continuous text.
Filler Words
Enable "Remove filler words" to eliminate common fillers (um, uh, hmm, etc.) plus any you add to the custom list. Context-dependent words like "like" and "so" are excluded by default to avoid removing legitimate usage. Detection is case-insensitive and whole-word only.
Smart Punctuation
Auto-captions often produce double periods, inconsistent spacing, and raw double hyphens. Smart punctuation normalizes ellipses (...), converts -- to em dashes (—), fixes spacing after punctuation, handles multiple punctuation (!?), and protects URLs and email addresses from unwanted changes.
Step 1 — Upload or Paste Your Transcript
Paste your transcript directly into the input box or upload a .txt or .srt file. All processing happens client-side in your browser—nothing is sent to a server. SRT files are detected automatically, and the tool will offer SRT export after cleaning.
If you have a YouTube video, download the auto-caption file from YouTube Studio. For interview recordings, grab the transcript from Rev, Descript, or Otter. The cleaner works with any plain text or SRT format.
Step 2 — Enable Cleanup Options
The Transcript Cleaner offers four main toggles: Remove timestamps, Remove speakers, Remove filler words, and Smart punctuation. Enable the options that match your input.
Remove Timestamps
Toggle this ON to strip all timestamp formats. The tool detects HH:MM:SS, YouTube-style (1:23), SRT time ranges (00:00:12,000 --> 00:00:15,500), inline brackets [00:12.34], casual formats (1m23s, 1h2m), and "at 1:23" patterns. One pass removes them all.
Example:
Before: [00:12.34] So in this tutorial 00:15 we're going to cover...
After: So in this tutorial we're going to cover...
Remove Speakers
Interview transcripts often include speaker labels like "Speaker 1:" or "John:" at the start of each line. Toggle this ON to remove the labels while keeping the dialogue intact. If you need to preserve who said what, leave this OFF and manually edit later.
Example:
Before: John: I think the key is consistency. Sarah: Absolutely, and...
After: I think the key is consistency. Absolutely, and...
Remove Filler Words
Enable this toggle to strip common fillers (um, uh, hmm, you know, I mean, etc.). The tool provides a default list and lets you add custom fillers in the "Additional fillers" field. Context-dependent words like "like" and "so" are excluded by default—add them manually if you want them removed. Filler detection is case-insensitive and matches whole words or phrases only.
Step 3 — Smart Punctuation Rules
Smart punctuation normalizes messy formatting that slips through auto-captioning. It runs automatically when enabled and handles five common issues:
- Ellipses: Converts multiple periods (.. or ....) to a single ellipsis (...).
- Em dashes: Replaces double hyphens (--) with proper em dashes (—).
- Spacing: Ensures single space after punctuation marks (periods, commas, colons, semicolons).
- Multiple punctuation: Preserves intentional combinations like !? or ?! while cleaning up accidental duplicates.
- URL/email protection: Leaves URLs and email addresses unchanged so links stay intact.
Example:
Before: So here's the thing..I think--you know--it works.Check out example.com/link
After: So here's the thing... I think—you know—it works. Check out example.com/link
Step 4 — Export Clean TXT or SRT
After cleaning, you have three options: copy the output to your clipboard, export as TXT, or export as SRT. TXT export is always available. SRT export is enabled only when your input is valid SRT—it preserves all timing data and cleans only the cue text.
TXT Export
Click "Export TXT" to download a plain text file. Use this for blog posts, scripts, or any workflow that doesn't need timing information.
SRT Export
If your input is SRT, the "Export SRT" button becomes available. The cleaned SRT file keeps all timecodes and cue numbers intact, so you can upload it directly to YouTube, Vimeo, or any video platform that accepts captions. Timestamps inside cue text are removed, but the SRT structure remains valid.
Common Mistakes & Fixes
- Over-aggressive filler removal → Review your custom filler list. Avoid adding context-dependent words like "like" or "so" unless you truly want them gone everywhere. The default list excludes these to preserve natural speech.
- Lost context from speaker removal → If your transcript is an interview with multiple voices, keep the "Remove speakers" toggle OFF. You can manually edit speaker labels later while preserving who said what.
- Broken SRT export → Ensure your input is valid SRT format. The tool will detect SRT automatically, but malformed files may fail. Test with a small sample first.
- URLs getting reformatted → Smart punctuation protects URLs and email addresses by default. If you see issues, verify the link format matches standard patterns (http://, https://, mailto:).
- Timestamps not removed → Some custom timestamp formats may not match the detection patterns. If a timestamp survives, note the format and use manual find-and-replace, or report it for future updates.
FAQs
- Do you upload my transcript?
- No. Cleaning happens entirely in your browser. Nothing is sent to our servers or stored anywhere.
- Which timestamp formats are removed?
- Standard formats (HH:MM, HH:MM:SS, inline [00:12.34]), SRT time-range/index lines, YouTube-style (1:23), casual formats (1m23s, 1h2m), and "at 1:23" patterns are all removed when enabled.
- Which file formats are supported?
- Paste text or upload .txt or .srt. The tool processes everything client-side. SRT export is available only when the input is valid SRT.
- How do filler words work?
- Toggle "Remove filler words" ON to remove common fillers (um, uh, hmm, etc.) and any you add to the custom list. Context-dependent words like "like" and "so" are excluded by default to avoid removing legitimate usage. Case-insensitive, whole words/phrases only.
- What does smart punctuation do?
- Normalizes ellipses (...), converts -- to em dashes (—), adds proper spacing after punctuation, handles multiple punctuation (!?), and protects URLs/emails from formatting changes.
- Can I export results?
- Export TXT always. Export SRT is available when the input is valid SRT; it preserves timings and cleans only cue text.
- Will this work for podcast transcripts from Rev or Descript?
- Yes. Paste any transcript format—timestamps and speaker labels will be detected and removed automatically. Toggle the options that match your input format.