Midjourney has recently rounded out its web-based image editing capabilities: you can not only edit images it generated itself, but also upload external images for partial repainting (inpainting). The new interface centralizes tools like Erase, Restore, and Expand Canvas under an “Edit” entry point, and adds a “Retexture” mode that lets you quickly swap materials and lighting while keeping the same composition. Below, organized by features and onboarding steps, we’ll clearly explain the key changes in this Midjourney update.
The web editing entry point now feels more “like software”
On the Midjourney web app, image cards now have an “Edit” button; clicking it takes you into a more complete editing interface. Common inpainting steps are now presented as explicit tools: use “Erase” to select an area, and “Restore” to undo or bring back parts of the mask. Compared with the more roundabout workflow before, Midjourney is now better suited for quick revisions.
External images can also go into Midjourney: upload and then inpaint
The most practical part of this update is that Midjourney now lets you upload images from your computer into the editor. After uploading, you can crop, expand the frame, and then use area selection plus text prompts to “add elements, remove bystanders, or change the background.” It’s recommended that your prompt clearly states “what to keep” and “what to change,” otherwise Midjourney may drag surrounding details along with the edit.
Retexture mode: keep the structure, swap materials and mood
Midjourney’s “Image Retexturing” is more like “keeping the shape skeleton and then re-skinning it”: it estimates the scene structure and then regenerates textures, materials, and lighting. In practice, it’s great for swapping materials in product shots (leather → metal), changing building facades, or creating a cohesive series from the same portrait. If you want the structure to stay more stable, keep the mask as tight as possible around the target area, and write fewer camera-change instructions and more about materials and lighting.
Expand canvas and regenerate: easier composition adjustments
In the editor, Midjourney supports enlarging the canvas by adjusting the ratio and aspect ratio, which makes it easy to “add negative space” or “fill in the background.” Related actions like “Transform, Enhance, and Regenerate” are also more centralized: first expand to create space, then regenerate to make the newly added areas match the original image’s style more closely. For posters or covers, this workflow is more stable than repeatedly repainting the entire image.
Compatibility with reference features and stricter moderation: limits to know before you start
This editor can be used together with personalization, style reference, character reference, image prompts, and other capabilities, making it suitable for building serialized visuals. At the same time, Midjourney is also testing a smarter V2 moderation system that checks the prompt, the input image, the mask, and the output simultaneously; if you get blocked, your first step should be to make the prompt more specific and reduce sensitive descriptions. Another practical point is that the rollout is still relatively cautious: some accounts need to meet certain usage conditions before the full entry point becomes visible.
