Turn a Hot Wheels Car Into a Life-Sized Supercar With AI

Step-by-step process

This project runs in three phases: generate the end frame, film your child, animate the gap between them. The order matters. Generate the end frame before filming so you know exactly what final result you're building toward.

Step 1: Photograph your background plate

Go outside and photograph the exact spot where the transformation will happen. This image becomes your background plate for the AI generation step.

What works: Get low. A 3/4 POV from knee height gives the most natural perspective for a car sitting on grass. The camera angle in the background photo needs to match the angle in your child's video.

What to avoid: Top-down shots. The AI struggles to place a car convincingly when the camera is looking straight down at the ground.

💡 Parent Insight: Take the photo on an overcast day if possible. Soft, diffused light is easier to match in the AI-generated image than harsh directional sun. It also means fewer lighting fixes in Step 3.

Step 2: Generate the life-sized end frame

Upload your garden photo to Nano Banana 2 (via the Gemini app) or ChatGPT Images 2.0. This step generates the final image — the full-sized car sitting in your garden as if it is actually there. This is the bookend that Kling will build toward in Step 5.

The prompt that worked:

Perspective Composite: Use the uploaded garden image as a strict background plate. Introduce a life-sized, monstrous supercar modeled precisely on the Dragon Blaster Hot Wheels car. Orientation: Front 3/4 POV facing forward. Details: Metallic teal-cyan body, translucent violet wings, gold-rimmed wheels. Lighting: Apply bright, high-key, diffused overcast daylight to match the garden. Cast a soft, natural ambient occlusion shadow on the turf. 8K texture, photorealistic.

Result: A photorealistic image of the full-sized Dragon Blaster in the garden. The shadow on the grass is what makes it convincing.

💡 Parent Insight: Describe your toy's specific colours in the prompt. "Metallic teal-cyan body, translucent violet wings, gold-rimmed wheels" kept the Dragon Blaster recognisable through the whole transition. Generic prompts like "red race car" produce something that bears no resemblance to the original toy.

Step 3: Check the lighting and fix it if needed

Compare your generated end frame to your garden photo. Check two things: the direction of the light and the colour temperature.

The first version of the Dragon Blaster came out with golden hour lighting — warm, directional, dramatic. The garden photo was overcast and cool-toned. The car looked like a sticker placed on top of the image.

If the lighting is wrong, add a conform instruction to your prompt and regenerate:

Lighting: Apply cool-toned, diffused overcast daylight. No warm highlights. Match the ambient occlusion of the background photo exactly.

This is the single most important quality check before moving to animation. A lighting mismatch in the end frame cannot be fixed in Kling.

Step 4: Film your child

Film 5-10 seconds of your child throwing the toy onto the grass. This becomes Frame 1 for Kling.

Two options for how to frame the shot:

• Third-person shot: Camera stays fixed, child is visible throwing the car. Easier to film, but the child is a spectator in the result.
• First-person POV: Camera held at chest height, angled down. The viewer sees the throw from the child's perspective. The child becomes the catalyst.

The first version used a third-person shot. Conor watched it and cried because he wasn't part of what was happening on screen. The POV version, where he throws the car and the camera follows it to the grass, made him feel like he triggered the transformation himself.

💡 Parent Insight: For children under 5, the first-person POV version lands completely differently. They watch themselves cause something impossible to happen. Film both angles if you have time. The POV edit is always the keeper.

Step 5: Animate in Kling 3.0

Upload Frame 1 (the toy-throw video clip) and Frame 2 (the life-sized end frame) to Kling 3.0. This is where the transformation happens.

Two prompts tested. Use whichever produces the cleaner result:

Simple version (often works best):

The boy throws the hot wheels car on the grass and it grows to be a full size car.

Mechanical version:

The hot wheels toy car in the first frame is thrown on the grass and then it mechanically builds into a full size version in the last frame.

Result: A smooth morphing transition from toy to full-sized car. The background stays stable because you've given Kling both endpoints. It doesn't have to guess where the animation is going.

💡 Parent Insight: Kling 3.0 handles plain language better than technical descriptions. "It grows to be full size" works because the model already understands physical growth. Adding technical jargon confused it and caused background flicker. Start simple, add detail only if the first attempt produces something unrecognisable.

Step 6: Review and show your child immediately

Watch the result before showing your child. Check three things: does the car stay recognisable through the transition, is the background stable, does the shadow appear on the grass at the end.

If all three are yes, show your child straight away. The reaction within the first 30 seconds of seeing it is the whole point of the project. Don't let it sit in drafts.