Verdict
Midjourney Editor is an editing surface inside the Midjourney subscription, not an independent image-editing SaaS. It is useful when an existing Midjourney image or an uploaded image needs generative changes that remain inside the Midjourney workflow: replace a selected area, expand the canvas, change composition, combine layers, or restyle the image with the current V8.2 Edit Model.
It does not replace a deterministic pixel editor. Photoshop is the stronger choice for precise masks, typography, color-managed production, or layered handoff; Firefly or Krea may make more sense when their editing controls, ecosystem, or privacy terms match the job better.
Search Jobs
- Midjourney Editor pricing and subscription access
- How Vary Region, Pan, Zoom, Remix, Smart Select, and layers differ
- How to edit an uploaded image in Midjourney
- Midjourney V8.2 Edit Model compatibility and limits
- Midjourney HD edit GPU cost
- Whether Editor uploads and results are public or private
Editor, Edit Model, and Discord Controls
The web Editor is the interface. The Edit Model is the V8.X model that interprets edit instructions and generates the changed image. Midjourney's Create and Organize pages open a lighter Editor for Midjourney generations; the full Edit tab is required for external uploads and advanced tools such as layers.
Discord controls such as Vary Region, Pan, Zoom Out, and Remix are earlier or narrower entry points into related operations. The web Editor brings inpainting, outpainting, movement, resizing, and masking into one canvas. It does not turn every older Discord workflow into the same feature: current Edit Model results are explicitly incompatible with Remix, while old Retexture remains available only by selecting V7 or earlier.
What Each Tool Solves
- Remix changes the prompt or parameters while creating a variation; current V8.X Edit Model results cannot be remixed.
- Erase / Vary Region masks an area so Midjourney regenerates only that part.
- Pan extends the canvas in a chosen direction, while Zoom Out expands context around the existing image.
- Move / Resize repositions, scales, rotates, or changes the canvas aspect ratio before regeneration.
- Smart Select builds a mask from positive and negative points, then erases the selection or the background.
- Layers combines uploaded images, supports ordering and local erasing, and flattens the composition when the edit is submitted.
- Retexture regenerates the whole visible image into a more cohesive style instead of changing only transparent areas.
The full Editor can download the generated image and a transparent PNG representing erased areas. Layered work is flattened on submission, so it is not a PSD-style nondestructive layer document.
V8.2 Edit Model Boundary
The current Edit Model supports Midjourney V8.1 and V8.2. It can modify an existing image from written instructions, create a composition from up to four reference images, and perform inpainting or outpainting in the Editor. It also works with Image Prompts, Style References, Moodboards, Personalization, and the HD parameter.
Current limitations include no tile parameter support and no Remix support for Edit Model results. The model replaced parts of the older Omni Reference and Character Reference workflow, but older gallery jobs can still require cleanup of legacy parameters before submission.
This is a fast-changing area: Midjourney's current V8.2 documentation should be checked again before documenting a long-lived team workflow.
HD Editing and GPU Time
Midjourney's GPU guide estimates an Edit Model prompt at roughly 1 Fast GPU minute in standard definition and 2.3 Fast GPU minutes in HD. HD therefore uses more than twice the Fast time of the standard edit. The version guide identifies standard output as 1024 pixels and HD output as 2048 pixels.
That makes HD a selective final-pass option rather than a default for every draft, especially on Basic. Standard, Pro, and Mega include Relax image generation, but queue behavior and feature eligibility should be confirmed for the exact edit workflow rather than assuming every Editor job is free in Relax mode.
Pricing Through the Parent Subscription
There is no separate Editor subscription. Access and generation capacity follow the parent Midjourney plan.
Basic costs USD 10/month and includes 3.3 Fast GPU hours per month.
Standard costs USD 30/month and includes 15 Fast GPU hours plus unlimited Relax image generations.
Pro costs USD 60/month and includes 30 Fast GPU hours, Stealth Mode, and unlimited Relax image and video generations.
Mega costs USD 120/month and includes 60 Fast GPU hours, Stealth Mode, and unlimited Relax image and video generations.
Extra Fast GPU time costs USD 4/hour on all four plans.
Annual subscriptions receive a documented 20% discount and are paid upfront. For Editor-heavy work, GPU time and privacy requirements usually matter more than the nominal entry price: Pro or Mega is required for Stealth Mode, while Standard is the first tier with unlimited Relax image generation.
External Images, Rights, and Moderation
Users must have the necessary rights to every uploaded image and remain responsible for what they input, create, and share. Midjourney applies its Community Guidelines to external-image editing, can block seemingly innocent prompts, and may suspend accounts for policy violations. Blocked jobs do not consume GPU time; GPU time is deducted when results are produced.
The rights question is therefore not solved by having an upload button. Do not upload client, celebrity, private-person, or third-party material unless the user has the necessary permission and the planned edit complies with Midjourney's terms and applicable law.
Visibility and Stealth Mode
V8.X Edit Model creations normally populate the Create and Organize pages and can appear publicly on midjourney.com unless Stealth Mode is enabled. The Editor guide says this applies to both Midjourney-generated and external images.
The same guide separately documents a private path for externally uploaded or multi-layer Editor work under the V7-and-earlier history behavior. Because those rules are version- and workflow-specific, sensitive work should not rely on a blanket assumption that every upload is private. Verify the current version, whether the job is submitted with the V8.X Edit Model, and whether Stealth Mode is enabled before using confidential material.
Best For
- Midjourney subscribers who want generative inpainting or outpainting without leaving the Midjourney ecosystem.
- Concept artists combining references, changing perspective, extending compositions, or exploring restyles.
- Creators who need fast visual iteration and can accept model interpretation instead of deterministic pixel control.
- Users whose plan includes enough GPU time for repeated edits, especially HD passes.
Avoid If
- The deliverable requires exact pixel edits, editable text, color-managed production, or a nondestructive layered file.
- Confidential images cannot be exposed to a public-by-default workflow and Pro or Mega with Stealth Mode is not acceptable.
- The user lacks rights to the source image or the intended manipulation conflicts with Midjourney's policies.
- A team needs an official public automation API or a repeatable high-volume editing pipeline.
- The task is basic cleanup that a conventional editor can complete more predictably without generative variation.
Sources & Verification
- Last verified: 2026-09-07
- Evaluation scope: Current Editor tools, V8.2 Edit Model compatibility, HD/GPU behavior, parent-plan pricing, external-image rules, and visibility boundaries.
- Method: Source-based research from the official product materials below. No hands-on testing was performed for this update.
- Official sources: