Creator Tech LabCreator tools, comparisons, and costs
Menu
Market
English

Decision guide · Design and thumbnails

Do you need Midjourney? Decide with one finished visual system

Midjourney is a subscription service for generating and editing images, and for turning a starting image into a short video. It is most useful before final layout: exploring visual directions, producing scene options and building a consistent source-image library. A useful first month should not end with a gallery of attractive experiments. It should end with one approved hero image, two channel variants, a reusable visual reference and files that survive the final publishing workflow.

Official sourceLast checked: 2026-09-02
Decision summary

Midjourney

Choose Midjourney when original visual direction is the bottleneck and you are prepared to select, correct and finish the result elsewhere. Start with Basic if a bounded paid test fits into 200 Fast GPU minutes. Choose Standard when repeated image exploration makes unlimited Relax generation valuable, and consider Pro for Stealth, more concurrency or Relax SD video. Under the terms, companies with annual revenue above USD 1 million, and their employees, need Pro or Mega to own generated assets. Do not subscribe merely for final typography, exact product representation or automatic campaign layout; pair the generated source image with a design or editing tool and judge cost per accepted asset, not prompts per month.

Good fit

Original scenes and visual direction

You regularly need distinctive backgrounds, concept frames or campaign imagery and can budget time for selection, rights checks, retouching and final layout.

Look elsewhere

Exact layouts, private briefs or automation

Your work depends on precise copy, faithful product details, confidential inputs, a public API or a repeatable unattended production pipeline.

Decision test

Complete one visual system twice

Finish a hero image plus two aspect-ratio variants, then repeat the process from a saved brief and measure corrections, GPU time and accepted outputs.

Start here

What is Midjourney? A visual exploration and generation workspace

Midjourney turns text and image references into generated images through its website and Discord, then provides variations, editing, organization, upscaling and short image-to-video tools. The current web experience can be the main workspace; Discord is no longer required for the basic create-and-organize loop. It does not replace the final production judgment that makes text accurate, products faithful, licenses clear and exports ready for a publishing platform.

Input

A brief, references and usage boundary

Define the subject, audience, format, visual constraints and rights of every uploaded reference before spending GPU time.

System

Prompt families and an approved direction

Keep prompt versions, seeds when useful, aspect ratios, moodboard codes and rejection reasons so later images follow a deliberate system.

Output

Selected source art, not a finished campaign

Export only approved images, inspect them at delivery size and finish typography, logos, product details and accessibility in the appropriate editor.

01

How the workflow should change

  1. Unbounded prompting

    Change several prompt ideas at once and save whichever image feels impressive in the moment.

    Controlled visual production

    Change one variable per round, record the accepted direction and explain why every finalist passes the brief.

  2. Unbounded prompting

    Use a generated image as the final thumbnail, poster or product visual without a delivery inspection.

    Controlled visual production

    Check anatomy, objects, text-like marks, edges and factual details, then complete layout and copy in a specialist tool.

  3. Unbounded prompting

    Count generations and assume a larger allowance automatically creates more usable work.

    Controlled visual production

    Track GPU minutes, editing time, rejected candidates and the cost of each asset that is actually approved.

Midjourney can help you

  • Explore several art directions from a written brief and combine text with owned or permitted image references.
  • Reuse visual intent through image prompts, style references, personalization profiles and versioned moodboards.
  • Edit regions, extend a canvas, upscale selected images and animate a still into a short SD or HD video.

Midjourney does not automatically

  • Guarantee exact lettering, a faithful branded product, factual accuracy, legal clearance or a platform-ready layout.
  • Keep ordinary-plan creations private by default or hide work posted in a shared Discord channel through Stealth alone.
  • Provide an unrestricted public API or authorize third-party automation, account sharing and resale of service access.
02

Three steps for a useful first paid test

  1. 01

    Lock the acceptance brief

    Choose one real deliverable, write what must remain accurate, list forbidden outcomes and gather only references you may upload.

  2. 02

    Run controlled rounds

    Begin in SD, alter one variable at a time, save rejection reasons and use references or a moodboard only after a direction earns approval.

  3. 03

    Finish outside the gallery

    Inspect the selected file at target size, correct it in the final editor, export the real deliverables and calculate cost per accepted asset.

Three Midjourney terms worth knowing
Fast GPU time
The monthly processing allowance used by Fast jobs. It resets on renewal, does not roll over and is different from elapsed waiting time.
Relax Mode
A slower queue with unlimited eligible generation on qualifying plans; images begin at Standard, while relaxed SD video begins at Pro.
Stealth Mode
A Pro and Mega visibility control for the Midjourney website. It does not conceal a creation from people in a shared Discord space.
01Price

Midjourney

PlanPriceBillingMarket
Basic$10.00monthlyUS
Basic$96.00annualUS
Standard$30.00monthlyUS
Standard$288.00annualUS
Pro$60.00monthlyUS
Pro$576.00annualUS
Mega$120.00monthlyUS
Mega$1,152.00annualUS

Last checked: 2026-08-31 · This result is an estimate. Check the official checkout before buying.

02Decision guide

01Use Midjourney before layout, not as the whole publishing stack

Its strongest role is producing source imagery and a visual direction; the final communication still needs deliberate assembly.

A creator usually arrives with a message, not a fully specified image. Midjourney can turn that message into several compositions, lighting choices and visual metaphors quickly enough to compare directions before committing to a shoot or illustration. The web Create page accepts text and uploaded images, while folders, search and the Organize page help keep generations available for later decisions. That makes it useful for thumbnail backgrounds, documentary scene concepts, course artwork, music covers and mood exploration where an original visual is harder to source than the final text treatment.

The handoff is the important boundary. A generated candidate can contain plausible but wrong objects, unstable small details or marks that resemble lettering. Treat it as source art. Select the candidate against a written brief, open it at target crop and size, then move into Canva, Adobe Express, Photoshop or another appropriate tool for exact titles, brand marks, product facts, safe areas and accessible contrast. If the entire job is arranging approved photos and accurate copy into known templates, a layout tool may solve the problem with fewer review steps.

  • Use a real publication brief with a target audience, message, crop, must-keep details and rejection criteria.
  • Keep the approved source image separate from the final file containing copy, logo, disclosure and platform-safe spacing.
  • Judge the second campaign made from the saved system, because the first can be fast without being repeatable.
Official evidenceCreating on WebMidjourney EditorImage Prompts

02Choose a plan by queue, privacy and accepted output volume

All four tiers are recurring subscriptions, and the useful unit is GPU time consumed per approved deliverable rather than a promised image count.

The verified US plan table lists Basic at USD 10 monthly or USD 96 paid yearly, Standard at USD 30 or USD 288 yearly, Pro at USD 60 or USD 576 yearly, and Mega at USD 120 or USD 1,152 yearly. Basic includes 3.3 hours, or 200 minutes, of Fast GPU time. Standard includes 15 hours, Pro 30 and Mega 60. Standard and higher add unlimited Relax image generation; Pro and Mega also add unlimited Relax SD video and Stealth. Extra Fast time is listed at USD 4 per hour. Annual billing gives the advertised 20% discount but charges the full year upfront.

There is currently no free trial on the website or in Discord, so the lowest-risk mainstream test is one month of Basic with a fixed acceptance task. The separate niji journey mobile app may offer a limited trial, but it is not a free Midjourney web plan. Basic is enough to learn whether the visual workflow produces usable assets. Standard becomes rational when slow, unlimited image exploration replaces repeated Fast purchases. Pro is not simply a quality upgrade: choose it when private website visibility, higher image and video concurrency, relaxed SD video or the over-USD-1-million commercial condition changes the decision.

  • Basic: bounded learning and occasional source images within 200 monthly Fast minutes, without Relax or Stealth.
  • Standard: recurring image exploration where unlimited Relax jobs matter more than immediate queue priority.
  • Pro or Mega: confidentiality controls, larger parallel workloads, relaxed SD video or the revenue-based rights requirement.
Official evidenceComparing Midjourney PlansFree TrialsGPU Speed: Fast, Relax, Turbo

03Build consistency from references, records and selective editing

Consistency comes from a controlled brief and preserved decisions, not from adding more adjectives to every new prompt.

The current default image model is V8.2. Model versions can interpret the same prompt differently, so record the version and important parameters when a direction is approved. An Image Prompt influences content, composition and color but is described as inspiration rather than an exact copy. Moodboards collect selected images into a reusable aesthetic direction and generate codes that change as the board changes. Save the exact code used for production instead of relying on a live board that may later be edited. Style references and personalization can support the system, but they do not remove the need to specify subject, context and deliverable.

Use the Editor after selection, not as an excuse to rescue every weak candidate. It can remix, inpaint, pan and zoom on generated or uploaded images, yet each operation consumes attention and may alter neighboring details. Change one important variable per round and keep a rejection log: wrong object, weak focal point, unusable crop, inconsistent character, unwanted pseudo-text or rights concern. When a candidate passes composition but fails one local detail, edit the region. When its basic communication fails, return to the brief instead of stacking corrective prompts.

  • Record model version, prompt, aspect ratio, reference type, moodboard code and the reason a candidate was accepted.
  • Use only references you own or may submit, and do not treat visual similarity as proof that an output is cleared.
  • Stop a round when the next correction costs more than regenerating from a clearer brief or finishing manually.
Official evidenceMidjourney VersionImage PromptsMoodboardsMidjourney Editor

04Budget resolution, upscaling and video before the final render

A compelling preview is not yet a deliverable; pixel dimensions, aspect ratio, motion length and finishing work determine whether it survives publication.

Official documentation describes V8.2 SD output at roughly 1024 by 1024 pixels for a square image and HD at roughly 2048 by 2048, with dimensions changing by aspect ratio. Subtle and Creative upscalers can double eligible SD dimensions, and an upscale may cost about two GPU minutes. The official planning table estimates an SD prompt at 0.8 GPU minutes and an HD prompt at 1.3, while edits and unusual jobs can differ. Those are capacity estimates, not accepted-image guarantees. Generate low-cost candidates first, reserve HD or upscaling for finalists and inspect the actual pixel dimensions rather than changing a DPI label and assuming more detail exists.

Video starts from an image as a five-second clip. It can be extended by four seconds up to four times, for a maximum of 21 seconds, and each extension costs the same GPU time as the initial generation. Every paid tier supports SD video in Fast mode; HD requires Standard or higher and remains Fast-only. The default batch of four SD videos is estimated at eight GPU minutes, compared with 26 for four HD videos, while a single-video batch can reduce spend. Use video for a tested shot or transition, then finish sound, captions, timing, disclosures and delivery encoding in a video editor.

  • Generate exploratory rounds in SD and spend HD or upscale time only after composition and usage rights pass review.
  • Choose video batch size deliberately; four alternatives consume more GPU time even when only one can be delivered.
  • Open the exported file at its real dimensions and test the final crop, typography, compression and platform preview.
Official evidenceGPU Speed: Fast, Relax, TurboImage Size and ResolutionMidjourney Video

05Treat commercial rights and visibility as two separate checks

A subscription can grant contractual asset rights without clearing third-party material, and ownership does not make an ordinary generation private.

Midjourney says subscribers own the images and videos they create to the fullest extent possible under applicable law, and ownership persists after cancellation. The important exceptions remain: third-party rights still apply; an image another member created remains theirs when you upscale it; and a company or employee of a company with more than USD 1 million in annual gross revenue needs Pro or Mega to own assets under the current terms. You are responsible for permissions in uploaded references and generated content. The terms also grant Midjourney a broad, perpetual license over submitted content and resulting assets, so confidential inputs require a policy review before upload.

The community is open by default. Ordinary creations can appear publicly and be remixable. Stealth is restricted to Pro and Mega and controls website visibility, but it does not hide work from people in a shared Discord channel. Use Stealth with the website Create page, direct messages or a private server when privacy is required, and confirm the setting before generation. Community rules apply to public and private work and prohibit unauthorized automation, third-party scripts, account sharing and resale of access. Midjourney generally does not offer a public API, so do not design an automated content factory around unofficial connectors. With the V8.X Edit Model, Editor results automatically appear in Create/Organize and are public unless Stealth is enabled; the older rule requiring an upscale applies to V7 and earlier. Turning on Stealth does not retroactively hide earlier images: their visibility must be changed separately.

  • Record the creator account, subscription tier, source-reference permission and final-use approval for each shipped asset.
  • Do not upload an unreleased product, client face or confidential brief until the organization accepts the platform terms and visibility model.
  • Use only supported interfaces and human-operated workflows; an unofficial automation shortcut can put the account and production history at risk.
Official evidenceUsing Images and Videos CommerciallyStealth ModeTerms of ServiceCommunity Guidelines

06Make the renewal decision from accepted assets, not gallery volume

The subscription is valuable only when approved source art saves more production effort than prompting, review, correction and finishing consume.

Use a simple monthly equation: subscription charge plus extra Fast time plus external editing or stock cost, divided by the number of assets that passed review and were actually used. If a USD 10 Basic month produces eight accepted hero images, the subscription portion is USD 1.25 per accepted image. If only two survive, it is USD 5. These figures exclude labor and are not performance claims; they show why hundreds of discarded generations do not lower the business cost. Track active review minutes and corrective edits beside GPU time so a cheap prompt does not hide an expensive approval loop.

At the end of the month, renew only if the saved visual system produces a second real campaign with fewer rejected rounds. Stay on Basic for occasional approved source art. Move to Standard when Relax image generation replaces extra Fast purchases or protects Fast time for deadlines. Move to Pro when Stealth or the contractual revenue threshold is necessary, not because the name sounds more professional. Cancel when licensed stock, a photographer, an illustrator, Canva or Adobe Express completes the actual deliverable with less uncertainty. Midjourney can be one strong source layer without becoming the center of the stack.

  • Count only images or clips that reached publication, client approval or an intentionally archived reusable library.
  • Separate subscription cost, extra GPU cost, human correction time and final-layout cost in the monthly review.
  • Save exports and production notes before cancellation; account cancellation does not automatically delete history, while full data deletion is a separate irreversible process after its reversal window.
Official evidenceComparing Midjourney PlansGPU Speed: Fast, Relax, TurboData Deletion and Privacy FAQ
03Frequently asked questions

Frequently asked questions

Is Midjourney free to use?

No free trial is currently available on the Midjourney website or in Discord. The official help page describes a separate limited trial in the niji journey mobile app. For the normal web workflow, plan a bounded month of Basic and define the deliverable before subscribing.

Do I need Discord to use Midjourney?

No. The Midjourney website provides creation, uploads, settings, editing, personalization, moodboards, organization and downloads. Discord remains an available interface, but a beginner can run the core production test on the web and avoid learning chat commands first.

Can I use Midjourney images commercially?

Midjourney says subscribers own the assets they create to the extent permitted by law, subject to its terms and third-party rights. A company with annual revenue above USD 1 million, or an employee of that company, needs Pro or Mega to own generated assets. Upscaling another member’s image does not transfer ownership.

Are my Midjourney images private?

Not by default. Midjourney is open by default, while Stealth is limited to Pro and Mega. Even with Stealth, work made in a shared Discord channel remains visible there, so private production should use an appropriate private surface and confirm visibility before submitting sensitive content.

Which Midjourney plan should a beginner choose?

Basic is the clearest first test when 200 Fast GPU minutes are enough for one defined visual system. Choose Standard after evidence shows unlimited Relax images remove a recurring limit. Choose Pro for Stealth, greater concurrency, relaxed SD video or the over-USD-1-million rights condition—not as a substitute for a clearer brief.

04

Comparisons

What to read next

Creator stack monthly cost
Official sources and checked dates15 sources