Video Generation

You can generate AI videos asynchronously by creating a task with zai.video.generations.create, passing parameters like prompt, size, fps, and duration, then polling zai.async.result.query until the final video URL is returned.

Yes, you can generate a video from an existing image by providing the image_url parameter in your video generation task, allowing you to create image-to-video content with optional transitions for product demos or marketing clips.

Yes, the z-ai-web-dev-sdk is required on your backend to handle asynchronous video generation tasks. It provides the necessary methods to create tasks and query results until the final video URL is ready.

Supported parameters include prompt, image_url, size, fps, duration, and with_audio. You use these to customize AI video generation tasks for marketing clips, educational content, or creative media output.

Asynchronous processing creates a video generation task returning a task ID, then you poll for results until completion. This returns a final video URL and enables scalable media generation across apps without blocking your backend.

Best use cases include building product demos, social media marketing clips, tutorials, and educational content. You automatically produce video output by submitting text prompts or images to generate creative media for apps and services.