Midjourney Reference Style (Sref V7) is live: lock in your aesthetic direction with a single image

Recently, Midjourney has made “Reference Style (Sref V7)” a default capability for V7 jobs: instead of repeatedly describing “who it looks like” or “what kind of texture,” you can simply provide one style image to quickly align the scene’s color palette, brushwork, and overall mood. For anyone creating a series of posters or keeping a consistent visual identity across an account, this update is extremely practical.

What exactly has Sref V7 changed?

In the past, achieving a stable style in Midjourney relied more on long prompts, repeated rerolls, and fine-tuning parameters. Sref V7 is more like a “style anchor”: treat a reference image you like as an aesthetic template, and Midjourney will prioritize matching its overall visual language during generation.

One thing to note: Sref V7 mainly operates at the “style level”; it’s not a substitute for replicating a specific person’s facial features or the structure of a particular object.

How to use it on Midjourney Web: drag-and-drop or via parameters

In Midjourney’s web prompt bar, drag an image directly into the “Style Reference” area to enable Sref V7; then enter your text description and generate. Another way is to add --sref URL at the end of your prompt (where the URL is the image link), which also calls the style reference.

Because Sref V7 is enabled by default for V7 jobs, you don’t need to toggle anything on—what matters is providing the right style image, clearly.

What it’s good for: from series covers to brand visuals

If you use Midjourney to produce “different images under the same theme”—such as Xiaohongshu 3×3 grids, a set of course covers, or multiple layouts for an event key visual—Sref V7 can significantly reduce the disconnect of “each image looking like it was made by a different author.” It also makes it easier to unify color and materials when creating e-commerce mood shots, game concept designs, or illustrated storyboards.

It’s also team-friendly: share a unified style reference image, and different members can more easily generate work along the same aesthetic track in Midjourney.

Practical tips and cautions

When choosing a style reference image, try to pick one with strong stylistic characteristics: clear color tone, lighting, and material cues will be more stable than a cluttered collage. You can also prepare two or three style images to rotate and test, observing how much Midjourney converges stylistically on the same copy.

Also, Sref V7 is a style reference, not a content reference; if you want specific elements to be more controllable, it’s still recommended to clearly specify the subject, scene, camera, and materials in the prompt, then let Midjourney unify the presentation.