Image and video playground
Submit generation tasks from the console, inspect results, and reconcile usage.
Prepare
Sign in and open the playground. Choose image or video, select a model from the current catalog, and prepare a key allowed to call it. Review the price and billing unit before submitting.
The playground calls models using the key you provide. Check that its owning account matches the account you intend to use; billing attribution follows the actual key.
Submit a task
Select a model, enter a prompt, and configure the size, quality, duration, or reference image fields offered by the interface. Supported combinations vary. Correct unsupported parameters instead of repeatedly submitting them.
Keep the task identifier and inspect queued, processing, successful, or failed states. Video results usually require waiting for completion. Avoid repeatedly clicking submit while waiting.
Retrieve results
Preview successful results and save the images or videos you need. Result URLs may expire and are not long-term storage. If refreshing requires a key, provide valid credentials for the task.
For failed or delayed tasks, record the identifier, model, time, and error, then inspect usage. A request timeout is different from task failure.
Integrate in code
Start with the image API guide. Confirm video submission paths, queries, and response fields for the selected line. Console task states are a presentation layer, not a universal contract for all video APIs.