Midjourney Image Editor Launch: A Web Guide to Erase/Redraw and Expand the Canvas

The Midjourney Image Editor brings “editing after generation” to the web. You can upload images or use already-generated works, and directly erase, restore, expand the canvas, and inpaint specific areas—making composition adjustments and detail touch-ups much easier.

What exactly has been updated in the Midjourney Image Editor

The core of this update is a brand-new external image editor: it supports uploading images from your computer, then cropping, expanding, and repainting—even adding or replacing elements within a scene. Editing no longer relies solely on “re-imagine,” but instead combines region selection with text prompts to control the scope of changes.

It also introduces an “image retexturing” approach: first estimate the scene structure, then reapply materials and textures so lighting and surface qualities change holistically. For users who want to quickly switch “material styles,” this takes fewer steps than repainting from scratch.

How to enter the editor on the web and start editing

In the web version, you’ll see an “Edit” entry point to start the editing flow; click it to enter the new interface. The Midjourney Image Editor supports processing specific regions—you can first mark the target area, then add the elements or descriptions you want to appear.

If you’re starting by uploading a local image, it’s recommended to set the canvas aspect ratio and subject placement first, then do local inpainting. This makes each change more controllable and helps preserve the original composition intent.

The three most useful actions: Erase, Restore, and Expand Canvas

The Midjourney Image Editor provides “Erase” and “Restore” tools: Erase clears areas you don’t want; Restore reverts or brings back details. In practice, it feels like prompt-driven retouching—great for removing passersby, changing props, or fixing backgrounds.

Expanding the canvas is better suited for adding blank space for posters, turning a vertical image into a horizontal one, or adding more environmental context. You can extend the edges by adjusting the ratio and aspect ratio, then use regional prompts to align the new parts with the original lighting and perspective.

How to pair it with Style Reference, Character Reference, and Personalization

The Midjourney Image Editor is compatible with multiple “reference systems”: style reference (--sref), character reference (cref), and personalization models (--p) can all work together. When creating a series, you can use style reference to lock in the aesthetic, character reference to maintain consistency, and then use the editor to refine the image.

Note that character reference typically requires placing the cref URL after the prompt to avoid command-parsing errors; you can adjust strength with cw, ranging from “reference only the face” to “reference even the clothing.” If you only want to change the outfit, lowering cw tends to be more stable.

Access and moderation: Why you might not see it yet

Because the feature is very new, Midjourney is rolling it out in phases to certain user groups first—such as high-volume creators or long-term subscribers. If you don’t see the Midjourney Image Editor entry yet, it’s usually not an account issue, but simply that you’re outside the current staged rollout.

At the same time, the platform is testing a smarter V2 moderation system that checks everything end-to-end—from prompts and input images to mask selections and final outputs. When editing, it’s recommended to avoid borderline descriptions and sensitive elements to reduce the chance of being blocked or having to redo work.