vawe-launch

Run the vawe-launch pipeline: lock a brief at Step 0, capture the brand from your product URL, remix a theme preset, get storyboard sign-off, author the scene JSON from recipes, then finalize with author-check, seam-check, judge, and make video to render the mp4.

Use make tts with your narration script to generate a local offline voiceover WAV plus a word-timing JSON, then wire them into the scene's audio field. Run make pace-from-vo to align each beat's start and duration with the spoken words.

Three steps are user-gated: Step 0 (locking the brief), Step 3 (approving the storyboard proposal), and Step 6 (approving the final render). All other steps run autonomously, with autonomous mode posting the storyboard as a heads-up.

The destination choice sets the aspect ratio: 16:9 for YouTube or embeds, 1:1 for X, LinkedIn, and Instagram, and 9:16 for Shorts and TikTok. The engine can render multiple aspect ratios from one scene source in a single pass.

Step 6 runs author-check (validation, critique, design-spec, copy, assets, intent inspection), make audit for layout and contrast, seam-check for transition flashes, a vision judge scoring every frame, and a ledger pass against repetitive output.