WebVTT from a transcript: file structure and subtitle checks
How to build a WebVTT file from a transcript: the WEBVTT header, timestamps, cues, timing checks and testing in a player. Includes an example and checklist.
VTT turns a transcript into timed subtitles
A transcript preserves the whole conversation. Subtitles present short pieces of it at the right moments. Begin by dividing the text into cues, syncing them and checking how they read, instead of simply renaming the file. WebVTT is a format for time-aligned text tracks in web video.
A platform may add restrictions around encoding, line length, positioning or file names. Check the requirements where you will upload the file. This guide covers the basic structure and editorial review. Compatibility depends on the target system.
Build the file and check it against the video
Prepare a verified transcript
Correct names, numbers and terminology against the recording. Remove words only under the agreed editorial rules. Subtitles can be verbatim or condensed, but changes to meaning or attribution must not be hidden.
Split the text into short cues
Set a start and end time for every cue. Avoid unnecessary overlaps, avoid leaving cues on screen after their line ends and do not fit a long paragraph onto one screen. Break the text at a natural point in its meaning.
Write the WebVTT structure
Put WEBVTT on the first line, then leave a blank line. For every cue use a dot before milliseconds and a timing arrow made of two hyphens and a greater-than sign. Put the cue text below the timing line and follow it with a blank line.
Check in the intended player
Open the video with the file and watch the beginning, middle, end and areas with fast dialogue. Check for lag, cues that disappear early, text covering important visuals and awkward line breaks. Record problems by timecode rather than noting “somewhere in the second half.”
Worked example: short subtitle cues
This is a fictional training example. The names, circumstances and outcomes are included solely to explain the method.
First open the recording. Then use the timecode to find the line and check the text against the audio.
The sample VTT can use two cues: opening the recording and checking the line by timecode. Leave a blank line between them. You can omit hours in a short clip as long as minutes, seconds and milliseconds are formatted consistently.
Example cue: 00:00:02.000 --> 00:00:04.600, followed by the line “First open the recording.” These are illustrative timings. Find the actual boundaries by watching the source video.
| Checks | What to look for | If something is wrong |
|---|---|---|
| Header | WEBVTT on the first line | Fix the beginning of the file |
| Timestamps | Move forward and match the speech | Listen to the section again |
| Cue | Readable within its display time | Split or shorten it |
| Player | The track loads and is readable | Check platform requirements |
VTT verification sheet
Use it before delivering the file. It distinguishes valid syntax from viewing quality.
File and version: Video and duration: Platform / player: Encoding: WEBVTT on the first line: yes / no First cue: Last cue: Timing overlaps: Cues that are too long: Names and numbers checked: Sections to rewatch: Upload result:
Check the file beyond its extension
- The file begins with WEBVTT and is saved as text.
- All timing intervals were checked against the video.
- Cues preserve conditions and negatives.
- The finished track was watched where it will be published.
A closer look
How is VTT different from SRT?
The formats differ in their headers, timestamp syntax and supported features. Convert the file structure and test the import when changing formats.
Do cues need numbers?
A cue identifier is optional in basic WebVTT. Follow the target system’s documentation if it requires a particular form.