Wan 3.0 Overview: Prompts, References and Credits

Input modes

Text + Image

Prompt-only and reference-driven generation

Creative loop

Prompt -> Preview

Iterate from history and saved inputs

Default model

Wan 3.0

Preselected in the online studio

Wan 3.0 is designed around a compact production loop: describe the shot, optionally anchor it with a reference image, choose video settings, generate, then reuse the strongest prompt and media combinations from history.

Model-facing workflow

The site opens the video studio with Wan 3.0 selected by default. This removes the model-picking step for search visitors and keeps the first action focused on the shot: subject, motion, lighting, camera movement, aspect ratio, duration, and reference media.

• Text-to-video prompts work best when they include subject, setting, action, camera language, mood, and pacing.
• Image-to-video prompts should describe how the still frame changes over time instead of only repeating what is already visible.
• History matters: successful prompts become reusable creative seeds for adjacent cuts, product variants, and social edits.

Prompt and reference image control

Reference images provide visual continuity, while the text prompt supplies motion and editorial direction. In practice, creators get stronger results when the uploaded frame is clean, the action is specific, and the camera instruction is not fighting the composition of the source image.

Credit-aware iteration

Wan 3.0 is configured for conservative free usage: small welcome credits, controlled sign-in rewards, and paid plans for heavier experimentation. This keeps trial traffic sustainable while still letting new users test the core video loop.

The highest leverage improvement is not a longer prompt. It is a cleaner loop between prompt, preview, and revision.

Wan 3.0 production note