Unforgettable Car AI Prompt for the UNO 2 and 4 Scene

Assign Image 1 to the person and Image 2 to the car

Image 1 defines the face, skin tone, hair and visible clothing. Image 2 defines the vehicle’s make and model, body shape, paint, wheels, lights and trim. The reference video defines the glances, mouth movements, hand positions, card moment and final exit. Keep these references separate: the person photo must not define the vehicle, and the car photo must not define a face. Neither photo should replace the road or background.

The prepared direction follows one person through front-facing views, profiles and hands raised near the hair. It asks for the same face and outfit throughout. A multi-view sheet, if used, should show one consistent person rather than a group of possible cast members.

Open the Unforgettable Car AI studio · Follow the complete sunset-lot tutorial

Preserve the supplied card-to-car reveal timing

The supplied clip begins with a person, cards and an arm passing across the camera; the car then becomes visible behind them. The template uses Image 2 wherever that vehicle appears, preserving its source position and perspective through the car-only ending. Do not force the car into opening frames where it is not yet visible.

The scene direction also keeps the pavement, pink-orange sky, distant mountains, utility poles and wires, framing and camera movement. The replacement uses the reference car’s model and paint instead of preserving the source burgundy sports car. Lighting, shadows and reflections should adapt to the lot. Do not copy the vehicle photo’s daylight, background or camera angle, and do not introduce a new location or transition.

Review the card moment separately from face quality

The 2 and 4 are the visual reason for the “24” name. The prompt asks to keep those digits readable where the source shows them clearly, with natural fingers, contact and motion blur. A successful face replacement does not prove that the digits or hands are correct.

Pause on the card moment during review, then watch it at normal speed. Look for changing digits, a card merging into a hand or an extra finger. Keep the distinction between a naturally blurred moving frame and a persistent shape error. Repeating a render with identical inputs does not guarantee a fix.

Choose clear references for close views and the car-only ending

Use a clear face with visible eyes and hair, rather than a face covered by a hand or heavy sunglasses. Clothing in the photo helps define the outfit; unseen areas must be completed by the model. A distant full-body photo with only a few facial pixels can be less useful than a sharp upper-body view.

For Image 2, use one unobstructed whole car with wheels and both bumpers visible. Avoid collages with different cars or crops that hide the vehicle’s shape. The model must infer unseen sides, so details can vary; inspect the car-only ending for changes in body shape, lights or wheels. You do not need to copy the source gesture. The reference provides the hand movement and pose. Compare the resulting face in both directions when the person looks toward the car and back at the camera, then check the hair and outfit during the final turn.

Choose a portrait with enough facial detail

Use framing and audio controls for their actual jobs

A prompt is not the crop selector in this studio. Original preserves the complete 16:9 scene; 1:1 and 9:16 are centered crops that can remove the vehicle or cards. A missing car bumper caused by cropping should be diagnosed separately from a car that changes shape.

The provider receives silent video and the application restores the source audio afterward. Adding a song name to a prompt is not an available control here. The photo changes appearance, not the recorded voice or soundtrack.

Review preview and credit costs · Diagnose identity and motion problems