AI text-to-video generation: Wan and Seedance model guide
Text-to-video starts from a scene description instead of an uploaded frame. Define a subject, action and camera move, then choose the length, frame shape and sound you want in the resulting clip.
One free generation across all four tools. Create an account for 5 promotional credits. Model policies and results vary; review the settings and estimate before generating.
Available settings and credits
| Model | Resolutions | Length / output | Lowest settings |
|---|---|---|---|
| Wan 3.0 Pro Prime | 1080P, 2K, 4K | 2–30 seconds | 32 credits / 2s at 1080P |
| Seedance 2.0 | 480P, 720P, 1080P, 4K | 4–15 seconds | 18 credits / 4s at 480P |
These are current paid estimates at the lowest listed settings. Your form shows the exact cost for the configuration you choose. All four tools share one free generation per browser/device; signing up can grant 5 credits under the existing promotion.
Compare the available text-to-video models
Wan 3.0 Pro Prime offers fixed durations from 2 to 30 seconds and 1080p, 2K or 4K delivery. Its 2K and 4K tiers enlarge a native render that tops out at 1080p. Seedance 2.0 offers 4 to 15 seconds and 480p, 720p, 1080p or 4K, with native audio and an optional web-research step.
Standard Vidu Q3 is not offered here for text-to-video: the current provider endpoint requires a starting image. It is not silently replaced with another model.
Give the clip one clear action
Describe the opening scene, one main movement, the camera behavior, and how the shot ends. Keep the action achievable within the duration. For a short clip, a simple reveal or gentle push-in is easier to inspect than a montage with several scene changes.
Budget for length as well as resolution
Paid video credits depend on the selected model, resolution and seconds requested. Seedance web research can also change the rate. Every estimate comes from the current provider price card, using the same pricing rules as image-to-video. Review the displayed credits after each settings change before you queue the job.
Choose audio deliberately
Both available models support generated sound. Describe ambient sound, dialogue or music in the prompt when you need it, and disable audio for a silent visual draft. Audio availability does not guarantee perfect speech or synchronization; review the result before using it in a finished project.
A single queue for all four tools
Text-to-video jobs share the same queue as images, image edits and image-to-video. The queue shows position and progress, with an estimated ETA when the provider does not report one. You can move to another page and return to your browser or account history. For a specific approved starting composition, use image-to-video instead.
Start with the tool that fits your material
Use text-to-image for a new still, text-to-video for a new scene, image editing to revise an existing picture, or image-to-video to animate a starting frame. Compare image generation and video generation, and see credit packs before a paid job.