What problem does it solve? Building video generation into an application requires handling asynchronous task creation, status polling, and result retrieval, which is error-prone to implement from scratch. This Skill provides ready-to-use patterns for generating videos from text prompts or images with the z-ai-web-dev-sdk. ## Core Features & Use Cases - Text-to-Video and Image-to-Video: Create videos from text descriptions, single images, or start/end keyframe pairs, with base64-encoded image input support. - Asynchronous Task Management: Create generation tasks, poll for status with fixed intervals or exponential backoff, and extract video URLs from results. - CLI and SDK Workflows: Use the z-ai CLI for quick one-off generation or the SDK for batch processing, queue management, Express.js endpoints, and WebSocket status updates. - Use Case: Build a backend endpoint that accepts a product description, generates a 5-second promotional clip at 1920x1080, polls until completion, and returns the hosted video URL to the client. ## Quick Start Ask the AI to generate a 5-second video from the prompt "A cat playing with a ball" using the z-ai video CLI with automatic polling.
