Video Generation

To generate AI video from text prompts or images, you submit an asynchronous task using the supplied script and poll for the results to extract the final video URL. This automates multimedia creation with configurable size, frame rate, duration, and optional audio.

Asynchronous polling for text-to-video generation works by submitting a video creation task and then continuously checking its status until completion. This mechanism enables scalable batch workflows without blocking execution, ultimately returning a final video URL.

Yes, you need a backend environment and the z-ai-web-dev-sdk dependency to run image-to-video workflows. Backend execution is required to handle the asynchronous task polling and carefully process the results to extract the generated video URL.

Yes, you can configure custom size, fps, and duration when you create AI video from prompts. The generation workflow supports specific output dimensions like 1920x1080, adjustable frame rates, variable duration lengths, and optional audio tracks.

Limitations of asynchronous video generation include the necessity of continuous polling to retrieve results and the requirement of a backend environment with the z-ai-web-dev-sdk. You must carefully handle polling results to successfully extract the final video URL.