Skip to content
Transcribe Tools

SRT subtitle generator

Turn a recording into timed subtitles and fix the timing yourself before saving. Each subtitle has a start and an end you can edit, and the file is made on your device.

  • Free
  • No sign-up
  • No upload
Saves asSRTVTTTXTDOCX

Drop an audio or video file here

MP3, WAV, M4A, OGG, Opus, FLAC, WebM, MP4, MOV

What is inside an SRT file

SRT is plain text, which is why almost everything reads it. Each subtitle is a block of three parts followed by a blank line: a counter, a timing line, and one or two lines of text.

The timing line gives the moment the subtitle appears and the moment it disappears, as hours, minutes, seconds and milliseconds. SRT writes the milliseconds after a comma; WebVTT, the format web players use, writes them after a dot and begins the file with the word WEBVTT.

1
00:00:00,000 --> 00:00:03,200
This is a short sample.

2
00:00:03,200 --> 00:00:06,900
Ana: Each block has a number,
a timing line and the text.

Timing subtitles people can read

A transcript is accurate when the words are right. Subtitles also have to be readable in the time they are on screen. Broadcast style guides differ in detail but agree on the shape; Netflix’s English guide is a common reference.

  • Two lines at most, around 42 characters each. Break a long line in the text field and split it where a phrase ends.
  • Not too fast. About 20 characters per second is the ceiling for adult viewers; less for children. If a line is dense, extend its end into the pause that follows.
  • Not too short or too long. Under a second flashes by; beyond seven seconds viewers start re-reading.
  • Leave with the voice. End the subtitle when the speaker stops rather than when the next one starts. On this page the band under the waveform shows exactly where each subtitle sits against the sound.

Speakers and exports

Type a name in the speaker field of a line and it travels with the export: SRT and TXT write it before the text (“Ana: …”), VTT writes it as a voice tag that players can style, and DOCX sets it in bold. The speech model does not tell voices apart, so the names are yours to fill in.

To correct a subtitle file you already have, use the subtitle editor. To caption a video and preview the result over the picture, see video to text.

Questions and answers

What is an SRT file?

A plain text file of numbered subtitles. Each one has a counter, a line with the start and end time (hours:minutes:seconds,milliseconds --> the same for the end), the text, and a blank line after it.

What is the difference between SRT and VTT?

WebVTT starts with the line “WEBVTT”, uses a dot instead of a comma before the milliseconds and can carry a speaker name in a voice tag. Web players (the HTML video element) read VTT; most video editors and upload forms accept SRT. This page exports both from the same transcript.

Can I change when a subtitle starts and ends?

Yes. Every line has a start and an end field. Type a new time as minutes:seconds, and the band under the waveform moves with it. The end can never come before the start.

Why do some subtitles stay on screen during silence?

The model marks the end of a phrase where it stops hearing words, which can include a pause. Play the line, note where the speech really ends on the waveform, and shorten the end time.

Does the SRT keep accents and non-Latin scripts?

Yes. The file is saved as UTF-8, which is what YouTube, Vimeo and current video editors expect.

More tools on this site