# 2026-06-21 Demo Walkthrough

## 5-Minute Flow

1. Problem
   - Traditional lecture recordings are useful but hard to reuse as modular teaching content.
   - MentorAI turns a reference lesson into a structured, original teaching video workflow.

2. Input Evidence
   - Reference assets: `storage/research/v2EN9GVEeMc/`
   - Transcript: `storage/research/v2EN9GVEeMc/transcript.txt`
   - Analysis: `docs/research/v1-reference-liuxin-video-analysis.md`

3. Transformation
   - Rewritten teaching script inside `apps/api/src/scripts/produceChildhoodWonderVideo.ts`
   - Blackboard-style slide design.
   - Segment structure: original text, focus, board notes, narration, visual prompt.

4. Authorized Voice Path
   - Consent record: `storage/voice-rights/teacher-liu-authorized/consent.md`
   - Cleaned source audio: `storage/voice-rights/teacher-liu-authorized/cleaned-audio/reference-youtube-v2EN9GVEeMc-clean.wav`
   - Provider-neutral config: `storage/voice-rights/teacher-liu-authorized/voice-model.json`

5. Final Output
   - Play `storage/demo/2026-06-21/mentorai-childhood-wonder-demo.mp4`
   - Backup path: `storage/childhood-wonder-video/childhood-wonder-full.mp4`

6. Why It Matters
   - Same workflow can apply to another lesson:
     - reference analysis,
     - storyboard,
     - character design,
     - authorized voice,
     - generated slides,
     - final video.

## One-Sentence Pitch

MentorAI converts a teacher's authorized lecture style into reusable, storyboarded, blackboard-style teaching videos with auditable voice rights and a repeatable production pipeline.

## Backup Talking Point

If authorized voice rendering is not ready by demo day, show the Edge-TTS video and explain that the voice rights package and adapter are ready; the remaining provider step is replaceable because the video pipeline consumes per-segment MP3 files.