Zhipu AI Image and Video Generation Pro

Problem

Engineers building design assets, product demos, or short-form video often need to turn a text prompt into usable images, or animate a still image into a 6-second clip. Zhipu AI Image and Video Generation Pro wraps the Zhipu AI open platform APIs into a local skill, reducing manual request assembly, polling, and file download steps.

How It Works

• Text-to-image: supports glm-image, cogview-4, and cogview-3-flash, with sizes such as 1280x1280 and quality set to hd or standard.
• Image-to-video: uses CogVideoX-2 to generate about 6 seconds of video from a remote image URL or a local JPG, PNG, or WebP file.
• Watermark control: after signing the disclaimer and completing real-name verification on the Zhipu open platform, --no-watermark can be used.
• Prompts: specify subject, style, lighting, camera, or motion, such as a British Shorthair cat with green eyes, a slow push-in shot, or soft backlighting.

Boundaries

Set ZHIPU_API_KEY before use. Outputs may be blocked by content review, video generation often takes 60–90 seconds, and long jobs may need --timeout. Save local files and temporary URLs promptly.