Video tools need Sketch or above, for clips up to 15 seconds and 1080p. Canvas and above unlock longer clips and 4K. Free accounts may be offered a few free videos through the Welcome Pack.
Video
Text to Video, Frame to Video, sound, camera moves and timing.
- Text to Video creates a clip from a description of the scene and the action.
- Frame to Video animates a start frame you choose, and on supported models can move toward an end frame too. It is the best way to keep a look you already like.
At launch: Seedance 2.0 (the default), Seedance 2.5, Seedance 2.0 Mini, Kling 3.0 Omni, Veo 3.1, Wan 3.0, MiniMax H3 Max Turbo and PixVerse V6. Compare them on the Models page.
Video is priced per second of output, multiplied by the number of clips. Higher resolutions, premium modes and native audio cost more per second on some models. The Generate button shows the total before you start, and you are charged only for the seconds delivered.
Many models generate native audio: turn Audio on in the form. For any video, Add Sound Effect creates a soundtrack from a short description.
Pick a move — pan, tilt, zoom, tracking, follow, orbit, crane, jib, drone fly-through or FPV — and it is added to your prompt in a form the model understands. Presets are free and work best with Kling 3.0 Omni and Seedance.
Usually a few minutes, depending on the model, length and resolution. The tile shows elapsed time and an estimate. If a job runs far longer than normal it is stopped automatically and every credit is returned.
Yes, from its tile. If the model has not finished, the credits come back in full. If it finished just before you cancelled, the result is kept and charged.
Video and audio uploads are rolling out. Until they reach your account, tools that take a video or audio input accept clips from your Library, such as videos you generated.