AI Video Generator: Clips from Text or a Start Image, with Sound
Chat shortcut: /video
The AI video generator works in three steps: describe a scene or upload a start image, choose a model and settings, download the finished clip, with sound depending on the model.
With Account
Loading…
Short clips without a camera or editing software
A video generator turns a description or an image into a short clip with motion and light, and depending on the model also with sound. You don't need a location or actors: the scene is built entirely from what you enter.
You can choose between several models. They differ in maximum length, format, visual style and sound, so for important clips it is worth comparing them with the same description. A start image helps when a character, product or setting should stay as close as possible to your photo.
Typical uses are social media clips, short product videos, animated preview images or a mood video for a presentation. For explainer videos with a narrator, combine the clip afterwards with text to speech and captions.
How it works
- 1
Describe the scene or upload an image
Write what should happen, or upload a start image and describe the motion you want.
- 2
Choose a model and settings
Pick the model, length, aspect ratio, resolution and sound, as far as the model offers them; with a start image, some models keep its format.
- 3
Let it generate
Start the run and keep working; the finished clip appears in the chat. Expect a few minutes, the time varies by model and load.
- 4
Download and reuse
Save the video or edit it with the video tools, for example trim, compress or convert to GIF.
Typical uses
Social media clips
Reels or Shorts in portrait format from an idea, without a shoot.
Product videos
A product photo becomes a short clip that rotates or floats.
Animated preview images
Moving teasers for a blog, shop or newsletter.
Mood videos
Landscapes, cities or abstract scenes as a backdrop for presentations.
Storyboards
Make scene ideas for a commercial visible quickly before anything is filmed.
Good to know
- Input: a text description or a start image plus description; output: a video clip, with sound depending on the model and up to 30 seconds long (30 seconds currently only with Seedance 2.5).
- You set length, aspect ratio and resolution before starting, as far as the model offers them; the cost depends on model, length and resolution.
- Videos require an account with credit; this tool has no starting allowance for guests.
- Generation usually takes a few minutes and the time varies; you do not need to keep the page open.
- Finished clips stay in your history until you delete the chat and are not shown publicly.
- Readable text in the image, close-up hands and very fine movements are still difficult for video models; a second run costs tokens again. If a run fails, the reserved tokens are refunded.
Frequently asked questions
How long are the videos?
Depending on the model, a few seconds up to a maximum of 30 seconds; the 30 seconds are currently only available with Seedance 2.5. You choose length, format and resolution before generating, as far as the model offers them.
How long does generation take?
It varies by model and load, so expect a few minutes. You can keep working in the meantime; the video appears automatically in the chat.
Do I need an account?
Yes. For videos you need an account with credit, because generation is resource-intensive. The cost depends on model, length and resolution; if a run fails, you get the reserved tokens back.
Can I turn a photo into a video?
Yes. Upload a start image and describe what should move, for example "camera slowly pushes in, hair blowing in the wind". The image becomes the first frame of the clip; some models also keep its format.
Does the video have sound?
That depends on the model. With Seedance and Wan you switch sound on or off before starting; other models generate their own sound such as noises, ambience or speech. You see the setting before generating.
How do I write a good scene description?
Describe the subject, motion, camera movement, light and mood like a short brief. One action per clip works better than several scenes in one sentence.
Can I publish the videos?
You can use the clips for your own purposes. Mind the rights to any uploaded images, and keep in mind that platforms often require realistic AI videos to be labeled.