Translations (multilingual captions)
When a brand's content language differs from the spoken language, captions are translated automatically at render time:
- Translation happens per caption line, keeping each line's exact time span — so the rhythm of the captions matches the video in every language.
- Translations are cached per language and editable exactly like the original transcript (the Content tab shows one transcript per brand language).
- Editing the source transcript invalidates its translations (they re-translate on the next render) — they can never drift apart silently.
- The translator sees context: the clip's content brief (when one exists) and the surrounding lines, so terminology stays consistent across a long video.
- Requires your AI key (it's the one AI-dependent part of captions). If anything fails — no key, no connection — the render logs it and burns in the spoken language instead of failing.
