Review the detected chapters, choose which ones to render, and optionally override the voice per chapter before sending the job to the queue.
{% set analysis = pending.speaker_analysis or {} %} {% set analysis_speakers = analysis.get('speakers', {}) %} {% set show_analysis_prompt = pending.speaker_mode == 'multi' and not pending.analysis_requested %} {% set has_metadata = pending.metadata_tags %}No additional speakers met the threshold yet. All dialogue will use the narrator voice.
{% endif %}Press Analyze speakers after selecting chapters to discover recurring voices.
Add a phonetic spelling (IPA or plain text) so pronunciations sound right. Leave blank to use the written label.
{{ speaker.analysis_count }} detected lines · confidence {{ speaker.analysis_confidence|default('low')|title }}
{% endif %}