Synthesia dub speaker attribution vs Sume speech-to-text and audio

Synthesia now lets Enterprise users reassign, create and rename speakers in a dubbed video with one click. Sume has no speaker editor in the pages read here: it gives you audio detach, a speech-to-text route and audio joins, and the mapping of who said what stays in your own code.

Synthesia facts are from its updates page (entry dated 9/3/2026); Sume facts from Audio detach and Timeline audio, read 2026-10-01.

What did Synthesia change?

The entry says you can reassign, create and rename speakers in a dubbed video with one click, available on Enterprise plans. It is an in-product edit.

Which Sume pieces touch a dub?

Why detach first?

Detach returns a sample-exact wav by default, which the docs name as what Timeline audio and speech-to-text want. The source video is untouched. Poll GET /v1/jobs/:id/status and read GET /v1/jobs/:id/result; there is no per-resource GET for a detach.

Where do speakers get assigned?

The Sume pages cited here do not describe a speaker-attribution editor, so treat it as your own step: keep a table of time ranges and speaker names next to the transcript, and rebuild the audio from those ranges when a mapping changes. For a worked pipeline see build an AI dubbing pipeline.

What should I check before relying on this?

Read the speech-to-text docs for the exact response fields; this post only cites the route. If you need a one-click speaker editor inside the dub, that is what Synthesia describes.