A brief in, a film out. A language model plans four beats from your description and a look at your file, and lists what it said that you should check. The picture is your own: an STL is read in the browser, sorted into its separate bodies and opened up, and a photograph is given a depth map so the camera can move through it. The film plays by itself, on scroll, under the pointer or with the tilt of a phone, reports which beat was watched for how long, and downloads as one file with instructions an AI coding assistant can follow to put it in a site.
Optional. STL, PNG, JPEG or WebP. No file? A picture is made for you. Or try a sample:
It writes words. It does not draw the film: the picture comes from your own file, which is why the product on screen is your product and not an invention.