Narration that knows what it's saying.

AI narration in any voice.

• OpenAI multilingual TTS on Free; ElevenLabs voice library unlocks at paid tiers.
• Per-scene voice override when you want a specific character.
• Speed + tone controls per provider.
• Plan-gated minutes — generous on Pro+, no artificial throttling.

WhisperX word-level alignment.

• Every spoken word gets a millisecond timestamp on the timeline.
• Karaoke highlighting in the runtime — learners track along.
• Animations + visual cues fire on exact word boundaries.
• Language auto-detect; multilingual courses handled out of the box.

Cue-driven visual sync.

• `[show: revenue-chart]` and `[chart: q1-revenue]` markers in the script.
• AI cue-annotation pass adds them automatically.
• Layout AI uses cues to assign elements to slots and time their appearance.

Bring your own audio.

• Upload an MP3 / WAV per scene.
• WhisperX transcribes it + builds a narration script with sentence-level cues.
• No re-narration cost — your voice + the AI sync for free.

Worker-driven, never blocking.

• TTS + alignment run as BullMQ jobs.
• Authoring stays interactive while audio renders in the background.
• Per-scene retry on transient provider hiccups; the rest of the course keeps moving.