This article covers what's changed, how automatic generation works while you're editing, and what happens at publish time if anything is still missing.
What's changing
The “Generate Audio” button in the lesson editor has been removed. You no longer need to click anything to produce audio for a Read Aloud component — Kira now generates it for you in the background as you write.
How automatic generation works while editing
Audio generates automatically in the background whenever a Read Aloud component has text but no audio yet. This happens in two situations:
On mount — when a Read Aloud component loads and has text but no matching audio, generation kicks off right away.
On blur — when you click or tab out of a Read Aloud text field after entering or editing text, generation starts automatically.
You don't need to save, publish, or click anything else to trigger it — it happens on its own while you keep working.
Editing text that already has audio
If you go back and edit text in a Read Aloud component that already has audio, Kira only clears the existing audio once your edit causes the text to first diverge from what was originally generated. From that point, new audio is generated automatically the next time you click out of the field. This means small, unchanged re-visits to a field won't trigger unnecessary regeneration.
Status badges
Each Read Aloud component shows an inline badge so you always know where its audio stands. You'll see a badge while audio is:
Generating
Complete
Failed
This gives you a quick, at-a-glance way to check a lesson before moving on, without needing to click into each component.
The publish-time safety net
Because generation happens in the background, it's possible to publish or republish a lesson before every component has finished generating. To make sure lessons never go live with missing audio, Kira checks for this automatically at publish time.
If any Read Aloud components are still missing audio when you publish or republish, a progress modal appears and generates the remaining audio for you. Once everything finishes, the modal closes on its own and your publish continues — no need to go back and check components individually.
If some audio fails to generate
On the rare occasion that some audio fails to generate, you'll get a choice right in the same modal: retry generation for the failed components, or publish anyway and address it later. Either way, you stay in control of whether a lesson goes live.
Quick recap
The manual “Generate Audio” button is gone — audio now generates on its own.
Audio generates on component mount and on blur, whenever text exists but audio doesn't.
Editing text only clears audio the first time it diverges from what was generated, then regenerates on blur.
A status badge shows generating, complete, or failed for each component.
Publishing with missing audio triggers an automatic progress modal before the lesson goes live.
If audio fails, you can retry or choose to publish anyway.
Questions about Read Aloud audio in your lessons? Reach out to Kira Support and we're happy to help.