
Wan 2.7 Image-to-Image API by Alibaba
Edits and recomposes images with Wan 2.7 image using text instructions, multi-image references, and optional interaction boxes.
Alibaba WAN 2.7 Image Edit
Alibaba WAN 2.7 Image Edit enables instruction-driven image editing and multi-image composition. It is designed for workflows where you want to preserve the source scene, combine references, and steer the result with natural language.
Why creators love it
- Instruction-based editing: Rewrite clothing, materials, color, scene details, or object relationships with a plain-language prompt.
- Single-image and multi-image input: Works well for focused edits as well as reference-driven composition.
- Reference composition: Combine multiple images and a text instruction in one edit request.
- Practical controls: Supports
size,n,watermark,thinking_mode, andseed.
Perfect for
- Marketing teams adapting hero visuals for multiple campaigns.
- E-commerce teams refreshing product shots and styled composites.
- Users who want Wan 2.7 editing quality in the standard tier.
- Creative teams that need reliable visual revisions without rebuilding an image from scratch.
How to Use
- Provide a main source image, and add extra references when style, objects, or composition details need to be guided.
- Write a prompt that explains the desired edit clearly and specifically.
- Choose the target size and decide whether you want one result or multiple variations.
- Review the edited outputs and keep the one that best balances fidelity and creative change.
Pro tips
- Be explicit about what should change and what should remain stable.
- Use a single output when you want tighter control, and multiple outputs when exploring different directions.


















