Imaginatron creative tools
Cinematic AI prompts
A cinematic prompt gives the model a shot it can interpret. Describe the visible scene, where the camera sits and what the light is doing. For video, add the movement that happens during the shot. Use the worked example below to test those decisions in Create.
Account required to use the tool. Review current access requirements and any token quote in the app.
Write a shot brief with visible decisions
Start with a subject doing something in a place: a lone courier waiting at a rain-soaked station gives the model more direction than cinematic masterpiece. Establish framing next. A wide shot shows the empty platform; a close-up would concentrate on the courier's expression. Lighting should have a source, such as warm light from the station windows against a cool dawn sky. Lens descriptions can suggest a look, but they do not guarantee a particular physical camera or focal length. Leave out terms that compete with each other, such as locked-off camera and fast orbit in the same shot.

How to get started
Decide what the audience should notice
Write one sentence naming the subject, location and action. Choose a wide, medium or close view based on that purpose. Put the subject early in the prompt so the rest of the description supports it.
Choose light and composition
Name the light source and its direction. Describe what fills the foreground and background, and leave space where the action needs to happen. Use a short color description instead of a list of unrelated film styles.
Add motion for a video version
In Video mode, give the shot one camera move and a small amount of subject action. The example uses a slow move toward a waiting courier. For an image, select Image mode and replace the moving camera instruction with a still composition.
Test one revision at a time
Review the settings and token quote, then generate. If the platform is too busy, simplify the background in the next prompt. If the subject is lost, revise framing or contrast. Keep the successful wording so you can tell which edit helped.
An establishing-shot prompt to try in Video mode
Wide establishing shot of a lone courier in a dark green coat waiting on an empty railway platform at dawn. Warm station-window light falls across wet stone; the sky is cool blue. Railway lines lead toward the courier. The camera slowly pushes forward as a light breeze moves the coat hem. One continuous shot, restrained movement, no cuts or on-screen text.
Before you finish
Keep an image prompt focused on what a single frame can show. Video needs an action that unfolds over time. To continue a visual idea using a frame you already have, attach it in an image-to-video workflow and describe the movement instead of reinventing the scene.
Frequently asked questions
Which words make an AI prompt cinematic?
Useful wording describes framing, a light source and the relationship between subject and background. Cinematic alone leaves those choices open. A specific shot brief is easier to evaluate and revise.
Can I use the same prompt for an image and a video?
Reuse the subject, location and lighting. For a still image, remove time-based instructions such as a camera push. For video, explain the action and choose a duration that can contain it.
Should I put several shots into one prompt?
For an initial test, describe one shot. Generate additional shots separately and assemble them in the video editor, where you can control their order and timing.
Does the prompt set the output resolution or price?
No. Review the model and output settings in Create. Those options determine the available format and token quote; writing 4K or a camera name in the prompt does not set them.
For billing and account questions, visit the help center or read the full FAQ.