Skip to main content

AI Workbench: Transcriptions

Updated yesterday

Here you can transcribe and translate your videos.

The end users will see transcriptions and translations in the Player as closed captions tracks, and/or in the Transcript tab.

Transcription is a result of a speech-to-text AI processing and is only performed in the original language of the video. Usually your videos will be transcribed automatically during initial processing. If for any reason this processing was not performed, or the initial transcription was later deleted, you will be able to generate a transcription with AI in this part of the Workbench.

A transcription can then be translated into a wide variety of languages.

Translating a video will only translate the text assets for this video like closed captions and transcript (where available). The audio track and other contents of the video will remain unchanged.

A video can only have one transcription or translation in one language. So if a transcription or a translation in a language that you want to add already exists, it should be deleted before proceeding. You will be asked to confirm this action or to perform it manually depending on the flow.

There are the following control elements available in the Transcriptions screen:

[Translation by AI] will translate the transcript for this video into a language of your choice.


Languages dropdown: provides all existing transcripts and translations for this video. Select one to view.

Download action. Download a closed captions file or a transcript file. Closed captions file will have timestamps for every closed captions line.

Delete action. Delete the transcript and/or the closed captions in the currently selected language.

Edit Transcription action. Switch to Editing mode. There, user can edit transcription or translation.

Upload Translation action. Upload a closed captions (*.vtt) file in any language. It will become available in the Languages dropdown once uploaded. These closed captions will be available to your viewers in the Player.

Search box. Enter your search request and move between all instances where this item was found.

Autoscroll toggle. Control whether the transcript to the right is scrolled in sync with video played in the player to the left.

Highlight Speech toggle. Control whether the relevant part of the transcript is automatically highlighted in sync with video playback. This helps the user track which part of the transcript is currently being spoken.

The changes that you make in the Transcriptions part of the Workbench, like adding or removing a transcription or translation, are published automatically and affect the users the next time they load the video.

Edit Transcription or Translation

The original transcription and any translations can be edited directly in the Transcriptions screen.

To begin editing, the Edit button must be clicked. This switches the interface into Editing mode, allowing the user to make changes to the currently selected transcription or translation. The user can place the cursor on any word or segment to modify it and start typing.


While in Editing mode, hovering over a word will display a tooltip showing the word’s timestamp, a Play button to preview the corresponding audio fragment, and a Delete button to remove the word from the transcript. The user can also click directly into a word to modify it.

If the user removes an entire block of text - such as a word, phrase, or sentence - a Comment icon appears in its place. Hovering over this icon shows the original content in a tooltip, and the user has the option to restore the deleted text if needed.

Once editing is complete, the user can click Publish to save and apply the changes, or Cancel to discard them. Publish and Cancel buttons appear only in the Editing mode.

Did this answer your question?