Photo to cartoon video: animating a picture you already have

How image input actually works

Your picture becomes the opening frame of the video. Everything visible in it — the design, the palette, the composition, the lighting — is fixed at the start and tends to persist through the clip. That is the whole appeal: it removes the guesswork about what things will look like.

It also changes what your prompt should say. The image has already described appearance, so the text should describe what happens. "The fox turns its head toward the window and its ears twitch, then it stretches and settles down" is a good image prompt. "A cartoon fox in a forest" wastes the instruction, because you just described the picture you supplied.

Preparing the image

• Crop to your target shape first. The output inherits the aspect ratio of your image, and selecting 9:16 in the interface will not override it. If you want vertical video, crop the photo to vertical before uploading.
• Keep the subject large and clear. A character occupying a reasonable share of the frame animates far more reliably than one lost in a wide landscape.
• Use even lighting. A strongly coloured light in the source gets read as part of the subject's colour scheme and carries into every frame.
• JPEG, PNG or WebP. Ordinary photo formats, nothing exotic.

What this is genuinely good for

Bringing a drawing to life. Photograph a child's picture, upload it, and ask for gentle motion. This takes about two minutes and is reliably the most delightful thing you can do with the tool.

Protecting a brand asset. A mascot, a logo, a product shot — anything that must look exactly as it already looks. Describing these in text would give you an approximation; supplying the image gives you the real thing.

Rescuing a nearly-right generation. When text-to-video keeps producing the right idea in the wrong style, generate stills until one is right, then animate that. This is the standard professional workflow and it saves an enormous amount of prompt-rewriting.

Animating a photograph. A still portrait given thirty seconds of gentle motion is the most emotionally effective use of this mode, and the reason so many people try it first.

Controlling the ending too

Supplying a single image fixes where the clip starts. Supplying a second image as the last frame fixes where it ends, and the model animates the journey between them. That is how you land precisely on a logo, build a before-and-after transformation, or create a seamless loop by using the same picture as both frames.

Both modes are available in the cartoon video creator, and the trade-offs against text and reference input are compared in our guide to choosing an input.