Introducing Alibaba WAN 2.7 Image Edit on WaveSpeedAI
Alibaba WAN 2.7 Image Edit performs prompt-driven image editing with up to 9 reference images. Natural language control, composition preservation. REST API, $0.03 per run, no cold starts.
Alibaba WAN 2.7 Image Edit on WaveSpeedAI: Natural Language Image Editing With Multi-Reference Support
Describe the edit. Keep the subject. Change everything else. WAN 2.7 Image Edit transforms images using natural language instructions while preserving structural integrity - subject identity, pose, and composition stay intact while backgrounds, clothing, materials, and style transform according to your prompt.
How WAN 2.7 Image Edit Works
Upload 1-9 reference images and describe the edit in plain text. WAN 2.7 interprets what should change and what should stay, producing edits that feel intentional rather than destructive. Multi-image references enable fusion edits - combining elements from multiple sources into a coherent result.
Key Features of WAN 2.7 Image Edit
- Natural Language Control: Describe edits in plain English - no masks, no layers, no manual selections.
- Multi-Image References: Upload up to 9 images for style, subject, or background guidance.
- Composition Preservation: Maintains subject identity, pose, and scene structure during edits.
- Reproducible: Seed control for consistent iteration.
- Output up to 2048x2048: Production-ready resolution.
Best Use Cases for WAN 2.7 Image Edit
Fashion and Apparel
Swap clothing, colors, materials while preserving the model’s identity and pose.
Background Replacement
Change environment or mood without affecting the subject - studio to outdoor, day to night.
Product Variant Creation
Generate color/material variants from a single product photo.
Style Transfer
Apply artistic or visual shifts while maintaining structural integrity.
WAN 2.7 Image Edit Pricing
$0.03 per run (~33 runs per $1). For 4K output, use WAN 2.7 Image Edit Pro.
FAQ
What is WAN 2.7 Image Edit?
Alibaba’s prompt-driven image editing model with multi-reference support, composition preservation, and natural language control.
How many reference images can I use?
Up to 9 images per edit for style, subject, or background guidance.



