Photo to video prompts that work
A repeatable prompt pattern for photo-to-video renders: subject action, camera, light, and the continuity instruction that keeps the face intact.
The four-part pattern
Every prompt on this site follows the same order: one subject action, one camera instruction, one lighting note, and one continuity instruction. The action says what the person does. The camera says whether the frame moves. The light fixes the mood. The continuity line is what stops the model from re-imagining the face.
Example: “The person slowly turns their head toward the camera and smiles, then holds. Static camera, soft window light. Keep the face, hairstyle and clothing exactly as in the photo.”
Actions that render cleanly
Small, slow, single actions hold up best: a head turn, a breath, a blink, a smile forming, a hand raising to shoulder height, hair moving in a light breeze, a slow step toward the camera.
Fast or compound actions break down: dancing, running, spinning, anything that hides the face for several frames. If the body leaves the frame, the model has to invent it on the way back.
What to leave out
Do not ask for text, logos, brand names or a specific real person's voice. Do not describe violence, weapons or anything sexual. Those requests are refused at submission or produce a result that cannot be used.
Also leave out exact second-by-second choreography. The model responds to tone and direction, not to a shot list.
Start with the tool
Credits per motion are listed on the tool, packs start at $4.99.