defaultLanguage vs defaultAudioLanguage on a dubbed AI Short

Set both, and do not mix them up. In the YouTube Data API, snippet.defaultLanguage is the language of the text in the video's snippet.title and snippet.description, while snippet.defaultAudioLanguage is the language spoken in the video's default audio track. A Short with an English title and a Spanish voiceover has defaultLanguage English and defaultAudioLanguage Spanish.

Both definitions come from the Videos resource, read on 2026-10-03, and the videos.insert page lists snippet.defaultLanguage and localizations as settable on upload. Sume does not call YouTube, so these fields are yours to set; Sume's part is the audio and the burned captions.

What is the difference in one table?

The two fields describe two different things, and a dubbed Short is exactly the case where they differ.

Which one do I set on a Sume-voiced Short?

Look at the track you are about to upload, not at the script you started from. If you rendered a Spanish voiceover, the spoken language is Spanish no matter what language the title is in. If you also write a Spanish title and description, the text language is Spanish too, and the two fields agree.

Sume's side of the language question is narrower than YouTube's. In Video captions, language is only a speech-to-text hint (es, en and so on; omit it for automatic detection), and the docs say it never selects the style or the font. Video inspect takes the same kind of language_code hint when transcribe: true. Neither sets a YouTube field.

What does a body look like?

This videos.insert body has an English title over Spanish speech, with a Spanish localization added. Check each key against the page before you send it.

{
  "snippet": {
    "title": "Cold open, episode 1",
    "description": "A Spanish-language Short.",
    "defaultLanguage": "en",
    "defaultAudioLanguage": "es"
  },
  "localizations": {
    "es": {
      "title": "Apertura, episodio 1",
      "description": "Un Short en espanol"
    }
  },
  "status": {"privacyStatus": "private"}
}

What else should I check?

A dubbed Short still has to fit the 3-minute ceiling; measure the TTS before the render. If you want several audio languages on one video, read multi-language audio tracks, and for per-language text see the localized titles checklist. Burned-in captions are covered in auto captions and language limits.

What mistakes should I avoid?

The first mistake is setting both fields to the language of the script you wrote. If the voice track is in another language, defaultAudioLanguage should name the voice, because the Videos resource defines it as the language spoken in the default audio track.

The second is treating localizations as a substitute for either field. The insert page lists localizations.(key) with a title and description per language, so it carries translated text only. It does not change the spoken language of the file.

The third is forgetting that a Short with no speech has no spoken language to declare. A music-only cut made with a bed from Sume's music tools has nothing for defaultAudioLanguage to describe, and the pages I read do not say whether the field should then be omitted, so test one upload and read the result in Studio before you bulk-set it.

• Voice language goes in defaultAudioLanguage.
• Title and description language goes in defaultLanguage.
• Translated text goes in localizations.
• Burned captions are pixels and set none of these fields.