I Love the Smiths AI: A Familiar Elevator Scene, with You in It
medium.com
medium.com
The scope is deliberately small: one scene, fixed dialogue and sound, and no prompt box. You choose who wears the headphones and who starts the conversation. One photo works with a fictional co-star; two photos let you put both people into the scene.
The article walks through the result and its limitations. Hair and clothing details can drift, so the homepage shows a generated example alongside the input photos. Generation is paid, starting at $12.99.
I’d be interested to hear whether this kind of single-scene tool is useful enough on its own, or whether you’d expect it to be part of a broader video editor.