Write scripts and generate narration

Before you start, link to this section

• Generation requires an internet connection and your own supported provider API key. Provider usage can incur charges.
• The app supports configured narration providers including OpenAI, ElevenLabs, Google Gemini, and Azure Speech; available models and voices depend on the provider.

Step by step, link to this section

1. Open Settings → Narration, choose the Provider, Model, and Voice, and save its API key. For Azure, also set the resource Region.
2. Open a chapter and write its Narration script using spoken prose rather than copying every on-screen label.
3. Choose Generate narration. Listen to the result and edit the script if needed.
4. After changing a script, use Regenerate narration to replace out-of-date audio.
5. For several chapters or languages, open Options → Narration audio. Review the missing and outdated tracks, then run the batch generation action.
6. Translate narration scripts in Translations before generating alternate-language tracks. Test playback in the exported course.

What to expect, link to this section

Generated audio becomes a project asset associated with its chapter and language. Script changes can mark a track as out of date rather than silently regenerating it.

Tips & troubleshooting, link to this section

• The configured provider receives the script and generation request. Narration is an optional online workflow, unlike ordinary local editing.
• If generation fails, check the saved key, provider access, quota, model, voice, and Azure region where applicable.
• For a batch failure, review the failed tracks and retry those that remain missing or outdated.
• A provider key is saved locally with encryption; do not put it in course content or a public support message.