Wan2.7 Image

Generate images from long natural-language prompts and then edit them with written instructions, without exporting to a separate editor.

Enter exact Hex codes and their proportions, or pull a palette from a reference image, so every generation matches your brand or art direction.

Fine-tune bone structure, eye shape, brow arch, nose bridge and jawline to avoid the same template AI face across your characters.

Accepts up to 3,000 tokens of prompt text and renders dense typography, formulas and multilingual layouts with far better legibility than typical models.

Produce up to twelve style-consistent images per session, then select regions to move objects, change textures or remove distractions with pixel-level precision.