Generating
Once the words are approved, the Audio step has Generate audio. Sadorra makes every section from the approved words, in its approved voice. A short newsletter takes about a minute; the step shows how many sections are done while it works.
When it’s ready, the player plays the whole newsletter or one section at a time, and says that the voice is AI-generated.
If a voice is paused or revoked
A section whose voice is paused or revoked doesn’t play, even if its audio was already made. Audio from a call that was still running when the voice was stopped is thrown away.
Previews
Before approval, Preview (before approval) on the Voices step plays the start of a section in a narrator voice. Previews are never published.
The monthly allowance
Your workspace can generate a set number of characters a month (200,000 unless your plan says otherwise). Previews and brand-voice designs count. If a newsletter would go over, generating is refused with how many characters are left this month.
When generation fails
If the provider can’t make a section, the step says Failed and you can generate again. Nothing half-made is published.