The SRT that lands in your editor

Sample coming soon

Every cue above is real output of a Subtitles (SRT or VTT) job on the sample video, in the same rows the cue editor uses. Type in any of them: the caption on the video changes with you, the start and end times stay as they are, and Reset to original puts the subtitle file back. Nothing here is saved, and in your own job the same edit is one Save away from the download and the burned-in video.

What is an SRT

An SRT is a plain-text list of subtitle cues: a number, a start and end time, and one or two lines of text per cue. Editors and YouTube Studio read it; WebVTT is the same idea for web players.

What lands in your editor

The first three cues of the sample video as SRT, then the same three as WebVTT. Both come from one Subtitles job, unedited. Same cues either way: the WebVTT opens with its header line and puts milliseconds after a dot.

Edit cues before you download

Fix a name or a term, save, and the timing does not move. You can also upload an edited .srt, and Reset to original brings the generated text back.

Or burn them in

Choose Subtitles (Burned in) instead of Subtitles (SRT or VTT) and you get an MP4 with the captions rendered in: six fonts, four positions, your colours. Same cues, same editor before the burn.

Subtitle sources compared

Subtitle sources compared
FeatureYouTube Studio downloadExtensions and scrapersYoutubeToText
Videos it works onYour own uploadsVideos with captionsAny public or unlisted video
Where the text comes fromYouTube's auto captionsYouTube's caption trackGenerated from the audio
Cue shapeRolling two-line cuesAs YouTube stores themShort lines, cut at a pause
FormatsSRT, VTT, SBVVariesSRT and WebVTT
Edit before downloadIn Studio, own videos onlyNoCue editor, timing kept
Other languagesAuto translate, own videos onlyOnly tracks that existUp to 9 per job, same timing

How it works

  1. 1

    Choose Subtitles

    Add target languages now if you need them; each one adds the length of the video to the cost. Filler words are cleaned on every Subtitles job.

  2. 2

    Open the cue editor and fix any name

    Each cue shows its start and end time and its text. Retype a misspelled name and save. The times do not change. Reset to original undoes every edit.

  3. 3

    Download .srt or .vtt per language

    Drop the .srt on your timeline in Premiere, DaVinci or CapCut, or upload it to your own video in YouTube Studio under Subtitles. Use the .vtt for a web player.

Things worth knowing

The whole video, no watermark

Creators who fix the SRT from YouTube Studio report that a 15-minute track came into Premiere cut short at under five minutes. Here the subtitles run to the last cue of the video, and a video can be up to 12 hours long. There is no watermark in the SRT, the WebVTT or the burned MP4 on any plan, including the free minutes. What can stop short is a preview: on the free plan, a video longer than your remaining free minutes gives you the cues for that part and an Unlock button for the rest. Unlock covers this one video, or take a plan if you caption every week.

Proper nouns and niche vocabulary

Names, brands and technical terms are where auto captions go wrong most often: a podcast title turns into three unrelated words, a city name comes out as two syllables that are not a word. The subtitles here come from the audio through ElevenLabs Scribe v2, and they still deserve one pass for names before you ship. That pass happens in the cue editor, not in a text editor. Retype the word, save, and the timing does not move, so you cannot knock the track out of sync with a stray keystroke the way you can when you open an SRT in Notepad. The translations are made in the same job from the generated text, so a fix in the original does not carry over: each language has its own tab in the editor, and you correct the name there too.

Delivering to a client or proofreader

An editor on the Adobe forum makes several reels a week and wants an external proofreader to check the captions. The advice in that thread was to send the text and let the proofreader change only the words, never the timecodes. That is what the cue editor enforces: text is editable, times are not. Send the client the .srt for review, or paste the cues into a document if they do not open subtitle tracks. When they send corrections back, retype them cue by cue and download the final .srt. For the video that gets posted, run the same link as a Burned job: you get an MP4 with the captions rendered in, and you can restyle and re-burn a language for free within 7 days of the job. Deliver the SRT for the archive and the MP4 for the feed.

The problem in their words

I would upload the video to YouTube, then download the SRT file, re-upload the SRT file into Premiere, then just fix grammar mistakes.
Adobe Community, Premiere Pro discussions
I'd rather post one less Short than fix captions the way Studio wants me to on my phone.
A fitness creator, quoted by Cutup

What a typical 20-minute video costs

Free

10 min

Your first 10 minutes are free. They cover the start of a 20-minute video; the rest is locked until you unlock it.

Just this video

...

One payment for a 20-minute video, 20 minutes. No subscription.

Creator

... / month

600 minutes every month, the smallest plan that covers a 20-minute video.

The Deep Dive

...

600 minutes, one payment, credits never expire.

See all plans

FAQ

Questions.
Answered.

Can't find what you're looking for?
Contact us.

No. The audio of the video goes through ElevenLabs Scribe v2, and the word timings that come back are grouped into cues. YouTube's caption track is never read. That is why a video with no captions, or with auto captions switched off by the uploader, gives the same subtitles as any other video. The text is generated, so check names before you ship.
Every cue is one phrase or sentence: at most 42 characters, on one line or two, and it stays on screen between 1 and 7 seconds, with 2 to 4 seconds as the target. Cues do not overlap and run in order. The start is the first word's start time and the end is the last word's end time.
The auto-caption SRT from YouTube Studio uses a rolling two-line layout: the second line of one cue becomes the first line of the next. Premiere shows both, so every sentence appears twice. Creators on the Adobe forum work around it by re-saving the track in Studio. The subtitles here put one line in one cue, so nothing carries over.
Yes. Pick up to 9 extra languages when you start the job, 10 with the video's own language. Each language gets its own SRT and WebVTT with the same cue numbers and timecodes as the original. Every extra language adds the length of the video to the cost. A translation that fails is not charged. The translations are made from the generated original in the same job, so correct a name in each language's tab of the cue editor.
Up to 12 hours. On the free plan a video longer than your remaining free minutes gives a preview: the cues for the first minutes and an Unlock button for the rest. A preview is not offered when you add extra languages, because every translation would run on the full video. Add the languages after you unlock, or take a plan first.
Then choose Transcription instead of Subtitles. You get the spoken words as paragraphs with a timestamp on each, speaker labels, search and replace, a summary, and exports to Word and plain text. Cues are cut for reading on screen, so they make poor notes. For lectures and tutorials see the notes page; for speaker-labelled quotes see the interview page. YouTube video to notes

Updated September 2026

Ready when you are

Paste the video link

Get subtitlesFirst 10 minutes free. No credit card required.