YouTube auto dubbing: Publish manually, and a 12-language Sume swap

YouTube auto dubs for eligible creators, with a Publish manually option. For a held ad, 12 languages on a Sume timeline spine swap is 12 x $0.10 = $1.20.

4 min readSume
All posts

YouTube's automatic dubbing is on by default for eligible creators, and its help page tells you to tap Publish manually to review dubs before they go live. That covers a channel video. A paid holiday ad you control end to end is a different job: you pick the voices, so you can swap the audio spine in a Sume timeline render at $0.10 per ceil(output minute). For 12 languages of a 30 second ad that is 12 x $0.10 = $1.20 of render cost, before the cost of making each voice track.

Both routes can work. The question is which one gives you a review gate you trust.

What YouTube lists

From the YouTube Help page on automatic dubbing: English can be dubbed to Arabic, Bengali, Dutch, French, German, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Malayalam, Polish, Portuguese, Punjabi, Russian, Spanish, Tamil, Telugu and Ukrainian. Videos over 120 minutes, without speech, or with copyrighted material are ineligible, and YouTube warns that dubs can contain errors from mispronunciations, accents, dialects or background noise.

YouTube automatic dubbing facts (read 2026-10-04)
ItemWhat the page says
DefaultEnabled by default for eligible creators
ReviewTap Publish manually to review before publishing
IneligibleOver 120 minutes, no speech, copyrighted material
Turn offYouTube Studio app, Settings, Content, Automatic dubbing
QualityDubs may contain errors

The Sume spine swap

A timeline job takes one Sume-hosted audio file as the spine and a list of clips on it. For a localized ad, keep the same video[] slots and replace only audio.url and audio.duration_seconds per language. Each render is a separate job with its own Idempotency-Key, so a language key like ad-q4-de-v1 can be retried safely.

If one language is longer than the others, its duration_seconds changes the output length, and the last clip can only stop 0.5 s before the end of the spine, so check coverage.

Cost of the swap, 30 s ad (read 2026-10-04, timeline rate $0.10 per ceil minute)
LanguagesRendersRender cost
33$0.30
66$0.60
1212$1.20
2020$2.00

Pick by who reviews

Use YouTube's own dub with Publish manually for organic channel uploads. Use the spine swap for creative where the voice, script and legal text are signed off per market. In both cases a native speaker should hear the final cut before it runs.

  • Channel video: let YouTube dub, hold with Publish manually.
  • Paid ad: swap the spine, one job per language.
  • Never publish an unreviewed dub on a sale ad.

Making the 12 jobs safe to retry

Twelve languages means twelve submissions, and a flaky network will cause some retries. Build the key from the content, not the clock: market code plus a version you bump on purpose, such as ad-q4-es-v1. A retry with the same key and the same body returns the same job; a changed body under the same key is refused, so bump the version when you change the script or the clips.

Then poll each job at GET /v1/jobs/:id/status and read /result when it is ready. The result of a timeline render includes video_url, duration_seconds, segment_count and billable_minutes; log billable_minutes per language so the bill matches the table above. If any language voice track runs past 60 seconds, that render bills 2 minutes, which is $0.20, not $0.10.

Keep a short sign-off sheet per market: the script text, the reviewer, the date, and the artifact id. This is the equivalent of the Publish manually step on YouTube: a gate that a person passes before anything goes live.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume