Documentation

Add captions

Have Studio write captions from the speech in your video, then style and correct them.

Studio listens to your video and writes the captions, timed to each word. You choose the style and fix any word it misheard. About 2 minutes for a short video.

Before you start

You need an edit open in the editor, with a clip that has speech. You also need one of these:

  • Companion running on your machine. The captions are then made on your machine, at no cost.
  • An OpenAI API key connected in Settings → AI Provider. A ChatGPT subscription does not cover captions.

Add the captions

Press Captions above the timeline. You can also ask Astro: "Add captions to the whole video."

  • With a clip selected, Studio captions that clip.
  • With nothing selected, Studio captions every video and audio clip.

Studio sends only the sound of your clips, never the picture. The captions appear on their own layer.

The first time Companion makes captions, it downloads a voice model of about 490 MB. Later runs skip that step.

Choose a style

Select the captions on the timeline. Their settings open on the right.

SettingOptions
StyleClean, Bold, Karaoke, Pill, Neon, Typewriter
All capsOn or off
Text colourAny colour
Accent colourThe highlight of the spoken word, in Karaoke, Pill, and Neon
Box colourThe box behind the words, in Pill
Size and PositionHow large the captions are and where they sit

Correct a word

The Transcript under the style settings lists every word. Click a word to edit it and press Enter. Clear a word to delete it.

Verify

Play the video in the preview. The captions follow the speech, in the style you chose.

Languages

Studio detects the language. With Companion, captions work in 25 languages: Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hungarian, Italian, Latvian, Lithuanian, Maltese, Polish, Portuguese, Romanian, Russian, Slovak, Slovenian, Spanish, Swedish, and Ukrainian. For any other language, Studio uses your OpenAI API key.

Troubleshoot

"Please set up an OpenAI API key to use this feature."

No Companion is online and no OpenAI API key is connected. Open Companion on your machine, or add a key in Settings → AI Provider.

"This clip has no audio track."

The clip is silent. Select a clip that has sound.

"No speech lines up with these clips."

Studio found no words in the sound, or the speech falls outside the part of the clip you kept. Check that the trimmed clip still contains the speech.

"Audio is larger than the 25MB transcription limit."

The sound of the selected clips is too long for one request. Select one clip at a time and caption each.

"Media is still preparing — try again in a moment."

A clip you just added is still loading. Wait a few seconds and press Captions again.

AI agent or LLM? Read this page as markdown

On this page