Midjourney New Feature Quick Look: How to Use Re-Texturing and Smarter Moderation

Midjourney has recently made “editing images” much smoother: not only can you upload local images for expansion, cropping, and partial repainting, but it has also added a more practical “image re-texturing” mode, while testing a more granular V2 AI moderation system. Below, based on real hands-on use, we’ll clearly walk through where to find these new Midjourney features, how to use them, and what to watch out for.

1. External Image Editor: Upload an Image and Edit It Directly

This time, Midjourney’s core change is expanding the editing workflow from “only being able to operate on generated images” to “being able to upload an image from your computer and then edit it.” After entering the editor, you can outpaint (expand the canvas), crop, do partial repainting, and add or replace elements within the scene.

The workflow is simple: upload the image first, then use area selection (like a mask) to mark the region you want to change, and finally use a text prompt to describe “what to change it into.” You can edit the same image repeatedly many times—ideal for poster revisions, retouching assets, and creating multiple variants of e-commerce images.

2. Image Re-Texturing: Keep the Structure, Change the Material and Lighting Overall

If you want to “keep the composition unchanged but swap all the surface feel,” use Midjourney’s image re-texturing mode. It first estimates the shapes and structure of the original image, then re-applies textures, causing the material, surface details, and lighting mood to change together.

Practical tip: In your prompt, prioritize “materials and lighting,” such as metal/ceramic/fabric, matte/glossy, warm vs. cool color temperature, studio hard light or cinematic soft light. This makes it easier for Midjourney to separate “keeping the form” from “redoing the texture,” and the results are much more consistent.

3. Editor Connected with the Reference System: sref, cref, and Personalization All Work

Midjourney also made the editor compatible with existing reference capabilities. You can continue to use style reference (--sref), character reference (cref URL), and image prompts while editing. For projects that need a “consistent art style” or “character consistency,” this linkage is critical.

In addition, the material mentions that --sref can be mixed with personalized models (--p): use --p first to maintain your usual aesthetic, then use --sref to pull the work toward a specific, well-defined style—useful for making a series of images or brand visuals.

4. V2 AI Moderation System: More Detailed, Stricter, and More “Whole-Picture Aware”

Midjourney is testing a smarter V2 AI moderation system. It checks prompts, input images, masked regions, and the final output as a whole. In other words, it’s not only what you write that gets reviewed—where you “select what to change” is also included in the evaluation.

To reduce false positives, keep prompts as straightforward as possible when describing visual elements and intended use, and avoid vague borderline wording. For uploaded images, choose materials with clear rights and safe content. If you get blocked, try narrowing the edit area or switching to more neutral wording—this is often more effective than forcing keyword changes.

5. Rollout Rules and Access Thresholds: Why You Might Not See It Yet

Because these features are very new, Midjourney is not rolling them out to everyone in the first phase. The material indicates that priority access is given to users who have generated at least 10,000 images (including annual subscribers), as well as users who have continuously held a monthly subscription over the past 12 months.

If your account doesn’t have the entry point yet, keep an eye on whether Editor-related buttons or update notices appear on the web version. Also, organize your commonly used Midjourney reference parameters into templates—so when the feature opens to you, you can apply them right away and save yourself a lot of trial and error.