Generate or edit versatile 1K and 2K imagery with several references and a clean, practical parameter set.
Wan2.7 Image turns clear creative direction into more controllable image results.
Unified generation and follow-up editing
A portrait generated by the model becomes the source for a targeted refinement: only the cobalt scarf changes to silver-gray silk while identity, pose, jacket, crop, and light remain.
Controlled portrait detail
Face shape, beauty mark, asymmetrical bob, blue eyeliner, single silver earring, silk scarf, and canvas jacket each follow a separate instruction without becoming a generic template face.
Multi-reference fusion
The cobalt bottle from one reference, Jiangnan waterside setting from another, and indigo, cream, and red palette from a third merge into a single believable campaign scene.
Structured text and layout
The readable word FORM, the Chinese character 形, a desk lamp, geometric line graphic, and coral rule occupy distinct zones in one disciplined editorial grid.
Key capabilities
Why choose Wan2.7 Image.
Build a simple, repeatable workflow around the inputs and controls the model reliably supports.
Capability 01
Unified generation and follow-up editing
A portrait generated by the model becomes the source for a targeted refinement: only the cobalt scarf changes to silver-gray silk while identity, pose, jacket, crop, and light remain.
The first result becomes the edit source
One requested variable changes
Portrait identity survives the refinement
Before
Result
Capability 02
Controlled portrait detail
Face shape, beauty mark, asymmetrical bob, blue eyeliner, single silver earring, silk scarf, and canvas jacket each follow a separate instruction without becoming a generic template face.
Facial marker and hairstyle stay distinct
Accessories remain correctly assigned
Skin, silk, metal, and canvas read differently
Capability 03
Multi-reference fusion
The cobalt bottle from one reference, Jiangnan waterside setting from another, and indigo, cream, and red palette from a third merge into a single believable campaign scene.
Product identity comes from reference one
Architecture and atmosphere come from reference two
Palette remains traceable to reference three
References
Result
Capability 04
Structured text and layout
The readable word FORM, the Chinese character 形, a desk lamp, geometric line graphic, and coral rule occupy distinct zones in one disciplined editorial grid.
English and Chinese text remain readable
Product and graphic elements stay separated
The layout uses a clear visual hierarchy
Prompt examples
Start with prompts shaped for Wan2.7 Image.
These examples define subject, composition, light, and motion so you can replace the creative direction with your own.
Unified generation and follow-up editing
A portrait generated by the model becomes the source for a targeted refinement: only the cobalt scarf changes to silver-gray silk while identity, pose, jacket, crop, and light remain.
View prompt+
Keep the exact woman, face, hair, earring, cream jacket, pose, crop, background, and light. Change only the cobalt scarf to reflective silver-gray silk.
The cobalt bottle from one reference, Jiangnan waterside setting from another, and indigo, cream, and red palette from a third merge into a single believable campaign scene.
View prompt+
Use reference 1 as the exact bottle, reference 2 for Jiangnan bridge, wet stone, wood and mist, and reference 3 only for indigo, cream and red palette. No text, 16:9.
