Video
AI Video Generator
Describe the subject, the action, and the camera move. A 4-second 720p clip usually takes a minute or two.
Describe a scene with motion, then start the generation.
What the text-to-video clip includes
Describe a scene and get a short clip: 4 seconds, 720p. You choose the frame: 16:9 for a normal player, 9:16 for a vertical short, 1:1 for a square, or 4:3. The clip is generated from the description, not from a photo you upload.
Motion is the part that matters. “A silver bicycle rolls slowly along a wet street at night, neon reflecting in the puddles, camera tracking beside it” tells the model what moves and where the camera is. A still description produces a dull clip.
A 4-second clip often takes a minute or two. The page keeps the request open and shows progress. Speech or on-screen words, if the scene has any, follow the language of this page.
How to generate a short video from text
Write the subject and the action
Say who or what is in the scene and what they do during the four seconds. One action is easier to read than three.
Add the camera and the place
Tracking, a slow push-in, or a fixed frame changes the clip. Name the place and the light: wet street at night, kitchen in the morning, snowy path.
Choose the frame
16:9 is the usual landscape clip. 9:16 is the vertical frame people mean when they search for a short. 1:1 and 4:3 cover square and classic frames.
Wait for the clip
Submit and leave the page open. When the clip is ready it plays in the panel. If the service is busy or out of quota, the page says so instead of hanging silently.
Scenes that make a 4-second clip
A vertical clip for a short
Frame 9:16. “Steam rises off a bowl of noodles as chopsticks lift a bite, camera close and still.” One motion, one frame, easy to judge in four seconds.
A landscape tracking shot
Frame 16:9. A bicycle on a wet night street with the camera moving beside it. This is the usual text-to-video request: subject, motion, camera, light.
A product turn without extra claims
“A matte black bottle rotates once on a wooden table, soft daylight.” Do not ask the clip to show text you have not written, and do not expect a voiceover unless you described speech.
Tips for a clip that matches the prompt
- Describe one action. Four seconds cannot tell a long story.
- Say what moves. A list of objects with no verb becomes a near-still image.
- Pick 9:16 only when you need a vertical frame. The default look of a normal video is 16:9.
- Name the time of day and the weather if they matter to the picture.
- Keep any spoken or on-screen words in the language of this page, and keep them short.
Video generator questions
How long is the video?
Four seconds at 720p. It is a short clip for a test, a social cut, or a visual idea, not a full edit. Generation often takes a minute or two.
Can I make a vertical video for shorts?
Yes. Choose 9:16. Use 16:9 for a standard player, 1:1 for a square, and 4:3 for a classic frame. The description should still say what moves.
Do I upload a photo or a clip?
No. This tool is text to video. The scene, the action, and the camera all come from what you type.
Why did it fail or ask me to wait?
A busy generator, a rate limit, or not enough quota returns an error on the page. Wait and try again. A very vague prompt can also fail to produce a usable clip.
Can the clip include words?
If you ask for speech or on-screen text, it should use this page’s language. Four seconds and 720p are a poor place for a paragraph. Prefer a visual action.
What should a good prompt contain?
Subject, one action, place, light, and camera. “A dog shakes off rain under a porch light, fixed camera” is enough. At least 8 characters, at most 500.