Skip to content

Video transcript example: text, timestamps, SRT and VTT

One 37-second clip transcribed four ways, shown in full: plain text, timed lines, an SRT file and a VTT file. What each is for, and what changed on the way.

A video transcript is the spoken words of a video written down. It comes in two kinds: plain text you read or paste somewhere, and timed text, where each line carries the moment it is said. SRT and VTT are timed text in the file formats that video players and YouTube accept as subtitles.

Below is one real clip transcribed all four ways, so you can see exactly what each looks like before you need one.

The clip

A 37-second explainer about compound interest, made by us: a narrated MP4 with a script we wrote first, so we know every word that was said. You can watch the clip.

It was transcribed on October 31, 2026 by Whisper large v3, run through fal. The video to text tool today runs Whisper large v3 turbo through Groq, so a run of the same file now can differ by a word or a comma. The four formats are produced by the same code the tool uses.

1. Plain text

What you get from Copy on the Text tab, or Download .txt. One paragraph, no times:

Compound interest is interest on interest. Put $1,000 away at 5% a year, and after the first year you have $1,050. In the second year you earn 5% on $1,050, not on the original $1,000. That is $52.50. The difference looks tiny at first. After 10 years, the money has grown to about $1,629, and after 30 years to more than $4,300. So the two things that matter most are the rate and the time. Starting 10 years earlier often beats saving twice as much later.

Use it for notes, quotes, a blog draft, or pasting into a document.

2. Text with timestamps

What you get on the With timestamps tab. Each line starts at the minute and second it is spoken:

0:00  Compound interest is interest on interest.
0:02  Put $1,000 away at 5% a year, and after the first year you have $1,050.
0:09  In the second year you earn 5% on $1,050, not on the original $1,000.
0:14  That is $52.50.
0:17  The difference looks tiny at first.
0:20  After 10 years, the money has grown to about $1,629,
0:24  and after 30 years to more than $4,300.
0:29  So the two things that matter most are the rate and the time.
0:32  Starting 10 years earlier often beats saving twice as much later.

Use it to find a moment again, to quote with a time, or to write chapters for a YouTube description.

3. SRT file

Download .srt. The subtitle format most editors and YouTube Studio accept. A numbered block per line, start and end to the millisecond, a comma before the milliseconds:

1
00:00:00,301 --> 00:00:02,481
Compound interest is interest on interest.

2
00:00:02,901 --> 00:00:08,441
Put $1,000 away at 5% a year, and after the first year you have $1,050.

3
00:00:09,221 --> 00:00:14,381
In the second year you earn 5% on $1,050, not on the original $1,000.

The full file has nine blocks; the other six follow the same pattern.

4. VTT file

Download .vtt. The subtitle format web players read (the HTML video element's track). The same lines, with three visible differences: the file starts with WEBVTT, there are no block numbers, and the milliseconds follow a dot instead of a comma:

WEBVTT

00:00:00.301 --> 00:00:02.481
Compound interest is interest on interest.

00:00:02.901 --> 00:00:08.441
Put $1,000 away at 5% a year, and after the first year you have $1,050.

If a tool asks for one and you have the other, those three differences are all there is to convert.

What changed between the script and the transcript

We compared the transcript with the script we recorded, word by word.

  • Words: 0 wrong out of 111. Every word in the script came back.
  • Numbers: 12 written differently. The narrator said "one thousand dollars", "five percent" and "fifty two dollars and fifty cents"; the transcript wrote $1,000, 5% and $52.50. That is the model choosing digits, not a mistake, and it is usually what you want in notes. If you need the words exactly as spoken, for a script or a legal record, check the numbers by hand.
  • Line breaks follow pauses, not sentences. Line 6 ends mid-sentence ("about $1,629,") because the narrator paused there. In subtitles that is fine; for reading, the plain text joins it back.

This was a clean voice with no music. How accurate is AI transcription has our test of the same kind of clips with music, noise and a second voice underneath.

Which one to download

  • To read, quote or paste into a document: plain text (.txt).
  • To find a moment or write chapters: text with timestamps.
  • To upload captions to YouTube or a video editor: SRT.
  • To add a subtitle track to a video on a web page: VTT.

Get one for your own video

Upload an MP4, MOV or WebM on video to text, or MP4 to transcript, and the result offers all four. For a public TikTok, Reel or Short, paste the link on its page instead. Three runs a day are free without an account, up to 3 minutes each; after that it is 3 credits a minute. If what you need is captions for a video you made, the subtitle generator starts from the same transcript.