Create a short video from a written scene
Text to video creates a 5–10 second clip from a prompt, without uploading a reference image. Describe the subject, setting, action and camera to give the scene a clear direction.
Start without a photo
In the studio, select Text to video. Write a scene, choose a frame and duration, then generate. Image ownership confirmation applies to uploaded images; text-only creation does not require an image upload.
Choose portrait for vertical compositions, landscape for wide scenes or square for a compact frame. Leave space around the main subject and avoid asking for many simultaneous actions.
A useful scene description
Example: “One adult dancer in a rehearsal studio performs a slow side step and a gentle arm movement. Full body remains in frame. Static camera, even lighting, natural proportions.”
A short phrase such as “Korean girl group dancing” leaves the number of people, choreography, framing and camera unspecified. More detail can reduce ambiguity, although a prompt cannot guarantee correct hands, faces or body movement.
What to expect
Complex choreography, crowds and rapid camera movement are difficult. A single subject, moderate motion and a stable shot are better starting points. Generated footage can contain artifacts and should be reviewed before use.
Generation time varies with the queue, duration and local compute availability. Five seconds is the clip length, not a promise of a five-second wait.
Begin with a five-second take.
Photo or text. Landscape or portrait. Check the current price and node availability in the studio.
Start creating ↗