Interactive: try "Play story", the three guided views, or "Present". Built with /archify.
Flow's agent rewrote prompts and dropped my face. Switching to direct mode, after a free test image, fixed it.
"Kakis" became "Caucus", "try" doubled. Whisper transcripts flagged it, and I re-shot with simpler lines.
I reviewed by timestamp; Claude moved cards to eye level or bottom and synced each pop-up to the spoken word.
Garbled signs and baked-in captions were blurred, covered or cut out.
/marketing:draft-content HI Claude, for the class exercise today, we will need to: * create my own virtual influencer * needs to have its own voice, styling, personality, a unique factor that compels the audience to follow/or know more about him/her * a 15 secs 16:9 video of the influencer introducing himself in his voice and a CTA liner for viewers. Something like a sizzle reel. * background should be in a natural (real location) environment setting, filled with ambient noise and foley sounds (if needed). If you need more information, ask me questions (one-by one) in an interview style, until you have enough context to execute the task /goal definition of done: the video showcase an influencer that does not look AI-generated and viewers are captivated with the influencer, the content he/she's into and the video is well executed.