WAN 3.0 Video Generation
Start with one video. AI-assisted setup, reference media, models and downloads.
Let AI help with the setup.
Copy the instructions to your AI. It will check your tools, then help configure or implement the integration.
Confirm the model and cost before testing. Execution requires an AI tool that can run code or call external tools.Choose your starting point
| Your situation | Next step |
|---|---|
| You want AI assistance | Connect with AI |
| You need step-by-step instructions | Your first video |
| You have an application or workflow | WAN3 API |
| You already submitted a task | Tasks and downloads |
Create a video-group key in the video console. Use an image model with a reference image, or a video model for text-only generation. Start with 480p and 2 seconds after checking the cost.
This is documentation, not an online generation canvas. An AI tool that can run code can help set up a client.
Save your result
Persist the task ID, wait for completion, then immediately save the MP4. Cached videos last about 10 minutes after becoming ready; individual signed URLs last up to 5 minutes. Your client should download automatically.
For still-image generation, use the separate image documentation. Its endpoints and billing rules do not apply to video.