A caption that lands half a second late reads as a mistake. shortshort times every word of a short on its audio and draws it in one of four styles, in the same studio where you cut and export the MP4.
30 free credits, no cardUp to 2 GB per videoMP4 · 9:16 or 4:5
Every word of a short is timed on the audio in a second transcription pass, then shown in one of four styles: Spotlight, Impact, Karaoke or Minimal.
Word-timed captions in shortshort: every plan, Free included.
Facts checked against the product on .
The first transcription covers the whole video and is used to choose the shorts. Once the passages are chosen, they are transcribed again on their own so each word gets a timestamp from the audio, not from an average reading speed.
If the re-transcription of a passage fails to match, the timestamps from the first pass are kept and flagged as estimated in the edit. You can see which words are exact and which are approximate before you render.
Spotlight, the default, sets the text in uppercase and highlights the word being said, one after the other. Impact is bold uppercase with one highlighted word per line. Karaoke lights words up as they are spoken. Minimal keeps them quiet and readable with no effect.
Colour and font can be changed for each video on any plan. If you already have an SRT timed to the source, import it and the captions follow your text. On the Studio plan, the colour and font are saved once as the account's single brand preset.
Each one lights the words as they are said; only the look changes.
5 steps, all in the same shortshort studio.
A published transformation from the community, made with the same studio: the original on top, the shorts below, with the transcript of each short.
3 of 6 shorts shown · The Override That Saved Humanity · watch it in the community, with the transcript of each short
Demo footage: Tears of Steel © Blender Foundation, licensed CC BY 3.0. Cut, cropped and captioned by shortshort.
Captions are in the language spoken in the video, never translated: each short's words are transcribed by OpenAI's whisper-1, which OpenAI says supports 98 languages, with accuracy that varies by language, so check the words before rendering. shortshort does not dub the audio or add a second language track.
They are drawn into the picture: the export is one MP4 at 30 fps, 1080 x 1920 in 9:16 or 1080 x 1350 in 4:5, with the captions in the frame. On every plan the studio can also export the captions as an SRT file, and the edit plan with every timed word as JSON.
No. shortshort adds no automatic B-roll, music or on-screen titles: the picture is the one you uploaded, cropped to 9:16 or 4:5, with the captions drawn on it.
Yes, on every plan, per video. On the Free, Starter and Creator plans you set them each time; on the Studio plan the colour and font are saved as the account's one brand preset and preselected on each new video.
Yes. You can import an SRT file that is timed to the source video, and the captions follow it. Time it to the original recording, not to the short, since the short's bounds are taken from that timeline.
No. Word timings come from the transcription, which is part of the AI analysis, so they need an account: 30 free credits at sign-up, no card, 1 credit per minute of video. The editor and the MP4 export need that free account too.
Free to try with 30 credits, no card. Exported shorts are listed in the public community gallery by default; any paid plan lets you keep them out of it.