Multiple model selection
Choose different video services based on picture style, action complexity, duration and budget.
Select the appropriate video model for each storyboard, and control the generation with the first frame, last frame, reference picture and action description. Long tasks are executed asynchronously, and failed shots can be retried individually without having to regenerate the entire episode for one shot.
A welcome gift for creators
1,500
creation credits
Your first story starts here.
Sign up for free credits to use on AI scripts, images, storyboards, and video.
One gift per new account. Added automatically.
Choose different video services based on picture style, action complexity, duration and budget.
Build status, error types, and results are saved in the project, and processing can continue after page refresh.
The video model generates each shot with native audio, followed by subtitles, shot review, and full-episode export.
Confirm the characters, scene, first frame, last frame and direction of action.
Set the video model, reference method and duration, and then generate the footage asynchronously.
Regenerate only the shots with problems, then export the full episode after subtitles are complete.
The time depends on the model, duration and current service queue. The task will run in the background, and the page will update its status through progressive polling, without the need to continuously maintain the request connection.
The system limits concurrency, uses a backoff strategy for queries that are safe to retry, and retains failed status for single-shot regeneration.
Yes. Your balance is checked before generation. Images and videos are billed for actual model usage, even if quality checks fail. The same call is never charged twice.