Seqvio

Seqvio turns real agent work, captured terminal/browser evidence, and authored ideas into narrated explainer videos. Use this production loop for new topics:

1. create and review EDITORIAL.md with seqvio-generate plan-editorial
2. create and review VISUAL-DESIGN.md with seqvio-generate plan-visual
3. use both approved artifacts with seqvio-generate plan-agent, then let the host agent produce ExplainerDocument with narration cues and ExplanationBeats authored together
4. validate and compile the IR to TSX with logical source timing
5. for long explainers, optionally run seqvio-generate render-plan and render chapters with --resume
6. extract and synthesize with seqvio-audio, which resolves phrase anchors and semantic scene time maps
7. run seqvio-qa with the resolved audio manifest, then render with seqvio-render