Caption files are simple text, but real ones are messy. Auto-generated files repeat lines. Subtitle tools leave tags like <i> or {\an8} in the text. Sound cues such as [music] or (applause) get in the way when you want to read the words. And if you cut the start of a video, every timestamp is now off. The cleaner in the Transcript section fixes these without any account or upload.
What the two formats look like
An SRT file is numbered blocks, each with a time range and the text:
1, then 00:00:01,000 --> 00:00:03,500, then the line of text.
A VTT file is similar but starts with the word WEBVTT, and uses a dot instead of a comma in the time: 00:00:01.000. Both are plain text, so you can open them in any text editor.
How to clean a caption file
- Open the cleaner and choose your file, or paste the text into the box. A .srt, .vtt or .txt file works.
- Choose what to remove:
- Style tags like <i> and {\an8}. On by default. - Bracketed cues like [music] and (applause). Off by default, because sometimes you want them. - Repeated lines. On by default. This merges lines that repeat across neighboring cues, which is common in auto-captions.
- Optionally set Shift timing in seconds. Use a positive number to move captions later and a negative number to move them earlier.
- Check the result in the output box.
- Copy the clean text, or download .txt, .srt or .vtt.
When to shift timing
Say you trimmed 4 seconds from the start of your video after making the captions. Enter -4 and every cue moves 4 seconds earlier. If the captions appear too early, use a positive number. Decimal values work, so 0.5 moves things by half a second. Always play the video once to confirm the sync.
Which output should you pick?
- .txt for reading, notes and writing.
- .srt for most editors and many video platforms.
- .vtt for web players.
If you plan to upload captions to your own YouTube video, YouTube's help page on adding subtitles and captions lists the file types it accepts and how to upload them. Check that page for the current list.
Fixing auto-captions properly
Cleaning is not the same as correcting. The tool removes noise, it does not fix wrong words. YouTube says automatic captions can misrepresent the spoken content and recommends reviewing them. If you care about accuracy, read through the cleaned text once and fix names and terms.
Privacy and limits
- The file is processed in your browser and never uploaded.
- If no timed captions are found in what you pasted, the tool says so. It expects SRT or VTT style timing lines such as
00:00:01,000 --> 00:00:03,000. - It does not translate captions or check them for accuracy.
- Very large files may be slow on an old phone or computer.
FAQ
What is the difference between SRT and VTT?
Both hold captions with timecodes. SRT is the older and simpler one. VTT is the web standard, begins with WEBVTT and uses dots in the timecodes.
How do I remove [music] from a caption file?
Tick the option to remove bracketed cues. It removes text in square brackets and parentheses such as [music] and (applause).
How do I fix captions that are out of sync?
Use the Shift timing box. A positive number delays all captions, a negative number moves them earlier.
Is my caption file uploaded anywhere?
No. The cleaner runs in your browser.
Can it convert SRT to VTT?
Yes. Load an SRT file and download the result as .vtt, or the other way around.