How to Create SRT and VTT Subtitles with AI in 2026

Publishing a video without subtitles in 2026 means leaving out 85% of viewers who watch content on mute on the subway, in the office or in bed. And creating subtitles by hand is still one of the most tedious tasks in the editing flow: timing each cue, breaking lines, translating to other languages. AI changes the equation: a well-made SRT or VTT file from a 20-minute video is now produced in under five.

This guide explains how to create SRT and VTT files with AI from any audio or video: technical differences between the two formats, code examples, tools, how to control sync and line breaks, how to translate subtitles into multiple languages while keeping timestamps, and how to load them into YouTube, Vimeo, Premiere and HTML5 players.

SRT vs VTT: technical differences

Both are plain text files that match phrases to timestamps, but they belong to different generations. SRT (SubRip Text) was born in 2000 as the output format of the SubRip program for extracting subtitles from DVDs. VTT (WebVTT) is the modern W3C standard, designed for HTML5 players and the semantic web.

When to use each format

When to choose SRT

• Video editing: Premiere Pro and DaVinci Resolve import it into the timeline as an editable subtitle track
• Desktop players: VLC, MPC-HC, MX Player auto-detect it if it shares the filename with the .mp4
• Uploading to YouTube and Vimeo: both accept it without conversion
• Client delivery: it's the format almost everyone knows how to open

When to choose VTT

• Your own HTML5 player: the <track> element of <video> only officially accepts VTT
• Courses and LMS platforms: Moodle, Canvas, Coursera or your own video player prefer VTT
• Styled subtitles: if you need colors, positioning or italics without burning text into the video
• Chapter tracks: VTT supports <track kind="chapters"> for marker-based navigation
• Modern web apps: React, Vue or any framework using the browser's native player

Internal structure of an SRT and a VTT

Looking at the file from the inside helps you understand how AI builds the result and how to fix it if something gets out of order.

.srt file example

1
00:00:00,000 --> 00:00:03,200
Welcome to today's podcast.

2
00:00:03,500 --> 00:00:07,800
We're going to talk about how
to create subtitles with AI.

3
00:00:08,000 --> 00:00:11,400
In five minutes you'll have
an SRT file ready to use.

Each cue has three parts: an order number, a time range with the --> arrow and comma as the decimal separator, and the subtitle text (one or two lines maximum). A blank line separates cues.

.vtt file example

WEBVTT

NOTE Subtitles generated by VOCAP

1
00:00:00.000 --> 00:00:03.200
Welcome to today's podcast.

2
00:00:03.500 --> 00:00:07.800 line:90% align:center
We're going to talk about how
to create subtitles with AI.

3
00:00:08.000 --> 00:00:11.400
<v Speaker1>In five minutes you'll have a VTT file ready to use.</v>

VTT requires the WEBVTT header as the first line, uses a dot as the decimal separator and allows extras: NOTE comments, cue positioning (line, align, position) and inline tags like <v Speaker> for speaker diarization.

Create SRT and VTT with VOCAP

VOCAP generates both formats in the same transcription process, with phrase-level timestamps and respecting recommended lengths.

Loading VTT in HTML5 with <track>

The native HTML5 player supports subtitles in a standard way thanks to the <track> element. It only accepts VTT.

<video controls width="720">
  <source src="podcast.mp4" type="video/mp4">

  <track
    label="English"
    kind="subtitles"
    srclang="en"
    src="podcast-en.vtt"
    default>

  <track
    label="Español"
    kind="subtitles"
    srclang="es"
    src="podcast-es.vtt">

  <track
    label="Chapters"
    kind="chapters"
    srclang="en"
    src="podcast-chapters.vtt">
</video>

The default attribute marks the track activated when the video loads. If you serve the HTML from one domain and the VTT from another (e.g. a CDN), remember to configure crossorigin="anonymous" on the <video> and the Access-Control-Allow-Origin headers on the VTT server.

Styling VTT subtitles with CSS

video::cue {
  background-color: rgba(0, 0, 0, 0.7);
  color: #ffeb3b;
  font-family: "Inter", sans-serif;
  font-size: 1.1em;
  text-shadow: 0 1px 2px #000;
}

video::cue(b) {
  color: #ff5252;
}

Only VTT supports this level of control. If you export SRT and need styling, you'll have to burn it into the video with ffmpeg or tools like HandBrake.

Uploading SRT/VTT to YouTube and Vimeo

YouTube

1. Go to YouTube Studio > Content > select your video
2. Subtitles tab in the left bar
3. Add language > pick the language of the file
4. Click Add under "Subtitles" > Upload file
5. Select "With timing" and upload the .srt or .vtt
6. YouTube activates them instantly; the CC button on the player shows them

YouTube also generates automatic subtitles in its own system, but quality in English is 88-92% and in Spanish 75-85%. Uploading your own SRT generated by VOCAP gives accuracy above 95% and improves indexing of the video in search.

Vimeo

1. Open the video in Vimeo and click Settings
2. Distribution tab > Subtitles section
3. Click + Add CC/Subtitles file
4. Upload the .srt or .vtt and select the language
5. Check "Available" so viewers can choose them

Importing SRT into Premiere and Final Cut

Premiere Pro

Since 2022, Premiere imports .srt files directly:

1. Window > Text > Captions > Import from SRT
2. Select the .srt file generated by VOCAP
3. A new subtitle track appears on the timeline
4. Each cue can be edited individually; drag the edges to adjust timing
5. To export the video with burned subtitles, in the Export panel enable "Burn captions into video"
6. To export as a separate sidecar subtitle file, choose "Create captions file"

Final Cut Pro

Final Cut prefers the iTT (iTunes Timed Text) format but accepts SRT with a workaround:

1. File > Import > Captions
2. Select the .srt; FCP converts it to iTT internally
3. The track appears on the timeline with editable cues
4. To export a CEA-608 or iTT track, use Share > Master File > Roles

DaVinci Resolve and CapCut

DaVinci Resolve imports SRT since version 18 (Edit > Import > Subtitles). CapCut Desktop and Web also support SRT since 2024 (timeline > Captions > Import file). On CapCut mobile, importing is more limited and it's better to generate subtitles from within the app from the audio.

Translating subtitles to other languages

The classic flow for translating subtitles was to run the SRT through a human translator or paste it cue by cue into DeepL. With AI, the process is reduced to a single step because VOCAP translates while keeping timestamps.

Best practices: length, timing, reading speed

The CSA (France), BBC (UK), Netflix Style Guide and CPL (Captioned Media Program in the US) guidelines agree on nearly everything.

Real use cases

Frequently asked questions