Build a scene from a text prompt
Choose text to video when you are inventing a scene. Describe one subject, one main action and a camera move the viewer can follow.

Write a scene, animate an image or build from visual references. Choose the starting point that gives your video the clearest direction.
An AI video generator creates a moving shot from a prompt or visual input. Choose the starting point that gives you control over the subject, composition and action you want to show.
Choose text to video when you are inventing a scene. Describe one subject, one main action and a camera move the viewer can follow.
Choose image to video when you already have the opening composition. Give the model a motion brief rather than repeating every visible detail.
Explore cinematic scale, animated characters and live-action motion before deciding what to make.

Guide the setting with concrete details: a rain-soaked conservatory, a yellow coat and small lights gathering in the dark. Give the scene one event to follow.

A person in a yellow coat explores a dark, rain-soaked glass conservatory. Small blue-green lights gather ahead and brighten as the person raises a hand. Keep one slow, grounded camera view.
An illustrated object can carry the action. Start with a small paper world, then describe the dragon rising from the open book.

An illustrated paper dragon rises from an open pop-up book on a sunlit desk. The paper landscape remains connected to the book as the dragon opens its wings. Keep the paper texture and a clear view of the small world.
Surprise is easier to read when the action and response belong together. A growing loaf and a startled little baker make one connected animated moment.

A friendly orange monster in a cozy bakery watches a loaf rise far above the oven. It looks surprised, then smiles at the growing bread. Keep the expressive character and the warm room visible.
Describe the object, its material and its movement. A reflective sphere travelling through a wet street creates a visual idea with scale, texture and a clear camera direction.

A large mirrored sphere floats through a wet city street. Reflections of buildings and lights move across its surface as the camera follows it at street level. Keep the scene grounded and the sphere readable.
An idea needs words. A defined composition needs a starting image. Visual references need a compatible model. Choose that input first, then give one shot a subject, an action, a camera direction and an ending.
Use applies these words to the generator. Choose the matching input mode and model settings there before generating.
Choose from the video models available in the generator. Input modes, duration and output settings vary by model.

Give a product a visual event that supports its shape and material. Let the reveal follow from the product itself.

Build surprise from a small action, then leave room for the character to react.

Build an idea into readable scenes for an early visual review.

Explore a product, setting and visual event before choosing your next direction.

Give an expressive character an action and a response.
Move between the available image and video tools as your project develops.
Use text for a new scene, an image for a defined opening, or references for a visual subject and setting.
Compare the available modes, then choose supported duration, resolution and aspect ratio in the generator.
Describe the main action and camera. Review the result, then change the part that matters before making another request.
Use your plan credits for images and videos. Review the selected tool, settings and credit quote before each generation.
$9.99/ month
800 credits / month
Credits renew monthly and expire at the end of the billing period.
View plan$24.99/ month
2,500 credits / month
Credits renew monthly and expire at the end of the billing period.
View plan$59.99/ month
6,200 credits / month
Credits renew monthly and expire at the end of the billing period.
View planUse text when the scene is still an idea. Use an image when the composition, subject or product is already defined. Reference mode is useful when the selected model accepts visual guidance beyond a single opening frame.
Some models accept multiple images; others use one starting image. Switch to reference-to-video and check the selected model’s image limit and input roles.
Describe one main action, what changes and how the camera moves. Replace abstract requests such as “make it cinematic” with a concrete subject, action and reveal.
Choose from the duration, resolution and aspect-ratio options displayed for the selected model. Different models support different settings.
An image can guide appearance, but small details and identity may still drift. Keep the action simple and inspect the result, especially faces, hands, labels and product geometry.
Completed generations appear in My Creations. Open a result there to view its status and the available download or follow-up options.
Choose an input, describe what should happen and set the format for your video.
Start creating