How to Use the Midjourney Editor for Remix, Vary, and Editing

The Midjourney Editor gives you control over specific parts of an image after generation. Instead of re-rolling the entire prompt or relying on blind variations, the Editor lets you target regions, change prompts mid-generation, extend the canvas, and retexture surfaces. These tools turn Midjourney from a one-shot generator into an iterative design environment.

This guide covers each Editor capability in detail: Remix mode for prompt swapping, Vary Region for inpainting, Pan and Zoom for canvas extension, and Retexture for surface replacement. Note that V8 Editor features may use V6.1 backends for certain operations, which affects output style consistency.

Remix mode: changing the prompt after generation

Remix mode lets you modify the text prompt when creating a variation. Normally, variations use the original prompt. With Remix enabled, Midjourney opens a prompt editor where you can change, add, or remove words before the variation is generated. This means you can keep a composition you like while changing the subject, style, or environment.

Enabling Remix

In Discord, type /settings and toggle Remix mode on. In the web interface, Remix is available in the variation options panel. Once enabled, every V1-V4 or Vary button will show a prompt editor before generating.

Practical Remix workflow

a woman in a red dress standing in a field of sunflowers, golden hour, editorial fashion --ar 3:4

After generating this image, select V2 with Remix on. Change the prompt to:

a woman in a blue evening gown standing in a field of lavender, twilight, editorial fashion --ar 3:4

The result keeps the general composition and pose from V2 but adapts the dress color, flower type, and time of day based on your new prompt. This is far more controlled than regenerating from scratch.

Remix tips

• Change one or two elements at a time for predictable results. Changing everything at once gives you essentially a new generation.
• Keep structural words the same (standing, field, editorial) to preserve composition.
• You can add or remove parameters during Remix. Want to try a different stylize value? Change it in the Remix editor.
• Remix works with Vary Subtle and Vary Strong. Combining Remix with Vary Subtle gives you targeted changes with minimal drift.

Vary Region: inpainting specific areas

Vary Region lets you select a portion of the image and regenerate only that area while keeping the rest unchanged. This is Midjourney's inpainting tool. You draw a selection over the part you want to change, optionally modify the prompt for that region, and Midjourney fills it in while matching the surrounding context.

How to use Vary Region

After upscaling an image, click the Vary Region button. A selection tool appears over the image. Draw a rectangle or freeform selection over the area you want to change. Then enter a prompt describing what should appear in that region. If you leave the prompt unchanged, Midjourney will regenerate the selected area in a way that fits the existing prompt and surrounding pixels.

Vary Region example

a modern living room with a large window, minimalist furniture, morning light --ar 16:9

After generating, the window view shows a cityscape but you want mountains. Select the window area with Vary Region and change the prompt to include 'mountain landscape visible through window' while keeping the rest. Midjourney regenerates only the window area with mountains that match the existing lighting and perspective.

Region selection tips

• Select slightly larger than the area you want to change. This gives Midjourney blending room at the edges.
• Small selections (less than 20% of the image) work best. Very large selections behave more like full regenerations.
• Be specific in your region prompt. If you select a table and want to change what is on it, describe the new objects clearly.
• Vary Region runs multiple times well. You can address different parts of the image in separate passes.

Pan: extending the canvas in a direction

Pan extends your image outward in a chosen direction: left, right, up, or down. The existing image stays in place and Midjourney generates new content that seamlessly continues the scene. This is useful for changing aspect ratios after the fact or revealing more of a scene that the original crop cut off.

Pan workflow

a narrow alleyway in tokyo at night, neon signs, rain-wet pavement, cinematic --ar 9:16

After generating this vertical image, pan left to reveal more of the street. The new content will match the existing rain, lighting, and neon style. Pan again to the left for even more extension, building a wider scene from the original narrow composition.

Pan considerations

• Each pan adds approximately 50% more canvas in the chosen direction.
• Content continuity is generally good for environments but may introduce inconsistencies with subjects (extra limbs, duplicated people).
• Pan with Remix enabled lets you change what appears in the extended area. Without Remix, Midjourney uses the original prompt.
• Multiple sequential pans in the same direction can produce diminishing coherence. Two pans works well; four pans in one direction often drifts.

Zoom: pulling the camera back

Zoom Out pulls the virtual camera back, revealing more of the scene around the existing image. The original image becomes a smaller portion of a larger composition. Midjourney generates the surrounding content based on the prompt and existing visual context.

Zoom levels

There are typically two zoom levels available: 1.5x and 2x. The 1.5x zoom adds moderate surrounding context while keeping the original image prominent. The 2x zoom doubles the canvas in all directions, making the original image roughly 25% of the total output.

close-up of a knight's gauntlet gripping a sword hilt, dramatic lighting, dark fantasy --ar 1:1

Zoom Out 2x on this result reveals the knight's full arm, torso, and potentially the surrounding battlefield. The original gauntlet stays as the focal point but now exists within a much larger scene.

Custom Zoom

Custom Zoom lets you specify a zoom factor and also change the prompt. This is one of the most powerful editing tools because you control both how much new content appears and what it should contain.

Custom Zoom prompt: a knight standing in a burning castle courtyard, dramatic lighting, dark fantasy --ar 16:9 --zoom 2

This extends the composition into a widescreen format while specifying that the surrounding environment should be a burning castle courtyard. The original gauntlet/sword image stays centered in the output.

Retexture: replacing surface materials

Retexture replaces the surface materials and textures in an image while keeping the underlying structure, shapes, and lighting intact. The geometry stays the same but the surface appearance changes based on your new prompt.

Retexture example

Original: a sports car on a mountain road, sunset lighting --ar 16:9
Retexture prompt: made of rusty metal and copper pipes, steampunk

The car keeps its shape, the road keeps its curves, the mountains stay in place, but every surface now has a rusty steampunk aesthetic. Retexture is essentially a style transfer that respects geometry.

Good uses for Retexture

• Converting realistic renders to illustrated or painted styles.
• Applying material studies (wood, metal, glass, fabric) to existing shapes.
• Creating style variations of a product design without changing the form factor.
• Turning photographs into stylized art while keeping proportions and perspective.

V8 Editor and backend behavior

In V8, some Editor operations may use V6.1 backends for processing. This means that certain edits, particularly Vary Region and Retexture, may produce output that has slightly different stylistic qualities than a pure V8 generation. The practical effect is usually subtle but can be noticeable in style-heavy images.

• Vary Region in V8 may process through V6.1, which handles inpainting with a different texture algorithm.
• Retexture backend selection varies. Check results and regenerate if style consistency matters.
• Pan and Zoom generally maintain V8 style better because they extend rather than replace content.
• If you notice style drift after an edit, try the same operation again or use Vary Subtle on the result to push it back toward V8 aesthetics.

Combining Editor tools in a workflow

The Editor tools work best when combined in sequence. A typical advanced workflow might look like this:

• Generate initial image with a well-crafted prompt.
• Use variations (Vary Subtle/Strong) to find the best base composition.
• Apply Vary Region to fix specific problem areas (wrong hand, artifact in background).
• Use Pan or Zoom to extend the composition if needed for a different use case.
• Use Retexture as a final creative pass if you want to explore alternative surface treatments.
• Upscale the final result.

Each step builds on the previous one. The key is to work from large decisions (composition, subject) down to small ones (texture, detail fixes), not the other way around.

Common mistakes

• Jumping into Vary Region or Remix before variations have produced a strong base composition.
• Expecting every Editor pass to look identical to a pure V8.2 generation. Some operations may use V6.1 backends and can shift texture or style slightly.
• Using Retexture when you need a pose, layout, or geometry change. Retexture keeps structure and only replaces surface treatment.
• Panning or zooming without a clear prompt for the new area, which invites filler that fights the original scene.
• Editing many regions in one pass so you cannot tell which change helped and which made the image worse.

Related guides

• For the variation and upscale workflow before editing, see /blog/how-to-improve-midjourney-results-with-variations-and-upscales.