What each part of the prompt does
- Overhead and locked
- A top-down view keeps the cup a clean circle so the pattern is easy to read, and a still camera lets all the motion come from the milk.
- One simple pattern
- A heart is the simplest latte art shape. Rosettas and swans need many precise movements that tend to come out as random swirls in 5 seconds.
- The hand kept small
- Only one hand and the jug enter the frame, which limits the chance of extra fingers or a warped grip.
- A payoff at the end
- The slow push-in on the finished heart gives the clip a final frame worth pausing on, which also makes a good thumbnail.
Variations to try
Same idea, one thing changed. These have no clip yet.
Iced matcha pour Side view close-up of a tall clear glass of ice and milk on a light wood counter. Motion: bright green matcha is poured in slowly and spreads through the milk in soft clouds around the ice. Camera: steady at glass height, slow push-in. Setting and light: bright morning daylight from a window behind the glass. Style: satisfying cafe video, crisp detail on the ice and swirls, fresh colours. Sound: ice clinking, a soft pour. Duration: 5 seconds, vertical 9:16.
Side view in slow motion Low side view of a wide white cup of espresso on a dark wood cafe counter. Motion: slow motion, a thin stream of steamed milk falls from a steel jug into the cup and the surface rises toward the rim. Camera: steady at cup height, very shallow depth of field. Setting and light: warm window light from behind, soft glow on the steam. Style: satisfying cafe video, crisp detail on the milk stream. Sound: the soft hiss of milk. Duration: 5 seconds, vertical 9:16.
If the result isn't right
- If the heart does not form, change it to "a round white circle of foam". It still looks good and renders far more often.
- Use a darker counter, such as black marble, for more contrast between the white cup and the background.
- Keep the cup wide and shallow. A tall glass hides the surface from an overhead camera.