Video Generation

You can generate video asynchronously by submitting a text prompt or image input to start a task, then polling the system to retrieve the final video URL. This handles long-running generation processes efficiently without blocking backend operations.

Yes, image-to-video generation supports both base64-encoded images and direct image URLs as input. You can provide these alongside optional parameters for frames, fps, size, duration, and audio configuration.

Yes, the z-ai-web-dev-sdk is a required dependency for executing AI video generation. You must install and configure this SDK in your backend environment to submit generation tasks and poll for results.

AI video generation allows you to configure video size, fps, duration, and optional audio. These parameters can be adjusted when submitting your text prompt or image input to control the final output specifications.

Task polling for asynchronous video creation involves continuously checking the status of a submitted generation task until completion. Once finished, the polling process retrieves the generated video URL for backend retrieval and usage.