Video Caption Generator
Upload a recording and get subtitles timed to what was said — as .srt, .vtt or plain text. Free to try, no account needed.
Runs entirely in this browser. Nothing you type is uploaded, stored or seen by us.
Three steps, no account.
-
01
Choose the recording
Audio or video. Choosing it starts it — there is no second button to press.
-
02
It is cut up in your browser
Your machine does the cutting and sends it a minute at a time, which is why an hour-long file works here at all rather than timing out on an upload.
-
03
Download .srt or .vtt
SRT for editors and most platforms, VTT for the web, or plain text if you only wanted the words.
Captions are not optional any more: most video on a feed is watched with the sound off, and a video with no captions is a video most people scroll past. This listens to a file that has never been captioned and writes the subtitle file for it, timed to the word rather than to a guess.
It is free to try without an account, with more on a free account and credits past that. Length is what it counts, because an hour of audio is not the same size of ask as a headline.
Video Caption Generator, answered.
Almost nothing, and getting it wrong breaks the file silently. SRT numbers every caption and separates the seconds from the milliseconds with a comma; VTT does neither and must begin with a WEBVTT line or it is rejected outright. Take SRT for a video editor or a social platform, VTT for a website video player.
Yes — it has to be, because that is where it is transcribed. Your browser cuts it up and sends it a minute at a time rather than uploading the whole file, which is why long recordings work, but the audio does reach us. We do not keep it after the transcript is made.
Export it as MP3. An MP3 can be cut up without being decoded first, so any length works; other formats have to be decoded in one piece and a long one will run a phone out of memory. Rather than crash half way through, it says so up front.
Because that is what is readable. Around forty characters a line and never more than two is the broadcast convention, and it exists so a caption fits a phone held sideways without covering the picture. Longer lines look more efficient and are harder to read.
Not from a link, no. If the video already has captions the YouTube transcript tool reads them instantly and for nothing, so try that first — this one is for recordings that have never been captioned.
This is one job. The workspace does the rest.
The free tools stay free and stay ungated. Recast is where they stop being separate tabs and start handing work to each other — a link becomes a card, an article and a week of posts without you moving anything between them.