The cheapest depth effect on the web: stack the photo, the text and a cut-out of the same photo. The person ends up in front of the type, the backdrop behind it, and a slow staggered entrance does the rest.
Nothing to attach. Both portrait layers are hosted here and the prompt already points at them, so the page runs the moment your AI hands it back. Try the live build ↗
The prompt
Paste this into Claude and open the file it gives you in Chrome. Nothing to install, no server.
Make it yours
Two swaps and it's you:
- The photos. Shoot on a plain backdrop, export the full frame and a background-removed copy at the same size. - The words. Your name in the marquee, your roles in the footer, your links in the header.
Re-run the prompt with your changes rather than editing the output by hand. The motion timings stay consistent that way.
