Create Kling 4.0 AI videos from a prompt, images, or video references. Choose 3–30 seconds and 720p or 1080p; see the credit estimate before you generate.
Clear references, deliberate shot control
Combine references, plan the action, and review the settings so the next iteration has a clear direction.
01 / START WITH A PROMPT OR REFERENCE
Describe the subject, action, and camera movement, then add image or video references when you want to guide the scene. Use the asset tags in your prompt to point to each reference.
02 / SET DURATION AND QUALITY
Pick a duration from 3 to 30 seconds and choose 720p or 1080p. Your credit estimate updates before you submit.
03 / DIRECT THE CAMERA
Write what the subject does, how the scene changes, and where the camera moves. Start with one main action so you can judge the result against a clear brief.
04 / REVIEW AND REFINE
Confirm the credit estimate, submit your task, and follow its progress in Create. Preview and download the finished video when it is ready.
From brief to finished clip
Set the shot goal, add optional image or video references, then check output settings and the credit estimate.
Flow AI VideoWrite the subject, action, camera move, and ending. When using references, name them in the prompt with their @tag.
Generate from text alone or upload images and video clips to guide the subject, setting, or visual style.
Choose a 3–30 second duration, 720p or 1080p output, and a supported aspect ratio. Review the credit estimate before submitting.
Open the Create workspace to follow progress, preview the completed video, and keep it in your Assets library.
One video, one clear purpose
Start from a specific shot brief for a product clip or social post.
E-commerce teams
Add product images as visual references, then describe material, action, and camera distance for a draft you can refine.
Prompt and shot inspiration
These project assets illustrate prompt directions. They are not verified outputs from Kling 4.0.
Flow AI VideoExample 1 / 3
Open on one useful detail, then reveal the full silhouette with a deliberate camera move.
Need a hand?
Understand the modes, reference inputs, available settings, and credit estimate.
Contact supportEnter a prompt to generate a video, or add images and video clips as references. This page currently uses the reference-to-video workflow.
The current API supports 3–30 second video and 720p or 1080p output. Available options are shown in the generator.
After adding an asset, use its tag, such as @image_1 or @video_1, to tell the model which visual reference to follow.
720p costs 20 credits per second and 1080p costs 30 credits per second. Choose a duration and resolution to see the estimate beside Generate.
Sign in, choose a duration, resolution, and model, then review the estimate beside Generate before submitting.
Bring the assets you already have
Enter a prompt and optionally add image or video references. Choose a format, duration, and resolution, then review the credit estimate before generating.
