Wan 2.7 unified image workflows

Wan 2.7 AI Image Generator

Use one prompt-led workflow for image creation or multi-image editing. Choose standard for economical 1K/2K output or Pro for higher fidelity and text-to-image output up to 4K.

Input

Model version
0 / 5000

For editing, state what should change and what must remain consistent across the source images.

Reference images (optional)

Upload up to 9 images for editing. Leave empty to create from text.

Aspect ratio
Resolution

Output

Your image will appear here

Write a prompt, optionally add reference images and choose the output resolution.

The model family, explained

What is Wan 2.7 Image?

Wan 2.7 Image is a unified model family for text-to-image generation and reference-based image editing.

Both versions accept up to nine optional source images. Standard supports 1K and 2K, while Pro adds 4K for text-to-image requests without references.

Each request runs as an asynchronous task and the completed output is stored privately for authenticated preview and download.

Creation and editing together

A flexible Wan 2.7 image workflow

Start from text or bring multiple source images into a single editing brief.

Bilingual instructions

Create or edit with focused Chinese or English prompts up to 5,000 characters.

Up to nine references

Combine multiple source images for product, subject and scene edits.

Standard or Pro quality

Choose economical 1K/2K output or Pro, including 4K text-to-image.

Tracked private output

Follow the task and download the completed image from private storage.

Three simple steps

How to create with Wan 2.7 Image

Choose a quality tier, write the instruction and review the output.

  1. 01

    Choose standard or Pro

    Use standard for economical work or Pro when final fidelity matters most.

  2. 02

    Write and add references

    Describe the desired image or edit, add up to nine references and choose resolution.

  3. 03

    Generate and download

    Submit the asynchronous task, then preview or download the privately stored image.

Wan 2.7 Image prompting tips

Editing prompts benefit from explicit preservation rules.

  • Separate the requested changes from the details that must stay untouched.
  • Use references with consistent subjects and a clear purpose for each image.
  • Remove references before choosing 4K Pro output, which is text-to-image only.
  • Keep one main editing objective per task for more predictable results.

Questions, answered

Wan 2.7 Image FAQ

Can Wan 2.7 create and edit images?

Yes. Leave references empty for text-to-image creation or upload source images for editing.

How many source images can I upload?

The model accepts up to nine reference images. FramePack currently accepts JPG, PNG or WebP files up to 30 MB each.

When is 4K available?

4K is available with Wan 2.7 Image Pro for text-to-image generation without reference images.

Can I download the result later?

The completed image is stored privately and remains available through the authenticated result experience.

Create or reshape an image

Start from text or up to nine references with Wan 2.7 Image.

Create with Wan 2.7 Image