Midjourney is a paid AI image generation service that creates images from text. You describe what you want in plain words and a few seconds later you get four images back, ready to enlarge, vary, edit or animate. Midjourney is an independent research lab, and its tool runs on its website, midjourney.com.
What you’ll take from this
This overview of Midjourney covers the essentials, with every fact checked against Midjourney’s official documentation in October 2026.
- Understand how this artificial intelligence tool works, without jargon.
- Know what you need to start today and what each plan costs.
- Write your first prompt with a clear structure and the basic parameters.
- Decide whether you can use the images commercially, and what risks to check first.
What is Midjourney AI used for?
Midjourney is an AI image generator that creates images based on text descriptions: illustrations, realistic photos, product concepts, concept art and other digital art. Since 2025 it can also turn an image into a short video. It belongs to the same family of generative artificial intelligence that lets ChatGPT write text, but this AI tool focuses on images.
Its strength is the aesthetic finish of the images, which is why artists and designers, art directors and illustrators pay attention to it. It is one of the tools covered in Founderz’s Spanish-language AI training for creatives.
Who is behind Midjourney
On its website, the lab says its own community funds it. Its leader is David Holz, co-founder of Leap Motion and the person who started the project, according to Wikipedia. The lab opened its public beta in July 2022 and, in its early years, you could only work through the official Midjourney bot on Discord.
How does Midjourney work? Inside the AI art generator
Midjourney works with an AI model trained to connect text with images. When you send your description, its machine learning algorithm reads your words and generates images that match your request. Each request, which the tool calls a job, produces a grid of four results.
What you can do with each result
From that grid you decide what to do with each output, as its guide to modifying your creations explains:
- Enlarge the image you like with Upscale to get an upscaled version.
- Create subtle or strong variations of one image with Vary.
- Edit one area, pan the frame or extend the canvas in the Editor.
- Turn the image into a 5-second video with Animate.
Which model version is the default
The model changes often, so check the date. V8.1 became the default model on June 10, 2026. V8.2 replaced it on July 24, 2026, according to the version documentation. V8.2, the default in October 2026, focuses on aesthetics, image quality and personalization. Older models remain. Pick V7 or V6.1 with the –v parameter when you need consistency with earlier work.
How do you use Midjourney to create images step by step?
To get started with Midjourney you need an account and an active subscription, because the tool generates nothing without a plan. The getting started guide describes four steps:
- Go to midjourney.com, sign in and choose a plan.
- Open the Create page, where you watch your images appear in real time.
- Type your description in the Imagine bar at the top of the page and press Enter.
- Wait for the progress to reach 100%, then pick which image to enlarge, vary or edit.
Everything you create is saved on the Organize page, where you can download, filter and sort your work into folders. Set up one folder per project on day one.
How do you write a good Midjourney prompt?
Good text prompts describe the subject first, then the setting and finally the visual style. Technical parameters go at the end. A specific, well-ordered description works better than a long list of adjectives.
A prompt template you can reuse
Use this structure as a starting point for any brief:
[main subject], [action or pose], [place and time of day], [style: photography, illustration, 3D render], [light and color], [framing or shot type] --ar 16:9
Here is the template applied to a specialty coffee shop campaign:
handmade ceramic coffee cup on a wooden table, steam rising, neighborhood coffee shop early in the morning, product photography, soft side window light and warm tones, close-up shot --ar 4:5 --no text, logos
On Discord, the /imagine command followed by the same text does the job of the Imagine bar, according to the Discord quick start.
Basic Midjourney parameters
Parameters go at the end, after two hyphens. These are the most useful at the start, according to the official parameter list:
| Parameter | What it does | Example |
|---|---|---|
| –ar | Changes the aspect ratio, which is square by default | –ar 16:9 for a website header |
| –no | Lists what to leave out | –no text, people |
| –v | Selects the model version | –v 7 to compare with an older model |
| –sref | Applies the style of a reference image, its colors, texture and light, without copying objects or people | –sref followed by the image URL |
| –chaos | Adds variety across the four results | Raise it while exploring ideas |
| –stylize | Controls how much artistic flair the model adds | Lower it if results drift from your request |
For a style reference, the Style Reference docs ask for text followed by –sref and the address of an image that is already online:
editorial portrait of a ceramicist in her studio, window light --sref https://example.com/reference.jpg
The –sw parameter, from 0 to 1000 with 100 as the default, sets how strong that style is. For tested starting points, Founderz has a collection of AI image prompt templates and a free image prompt generator.
Is Midjourney free? Plans and pricing
No, Midjourney is not free: you need one of its four paid plans, and its plans page offers no free trial. Midjourney offers plans that differ in fast hours, Relax mode, which generates without spending those hours, and Stealth mode, which hides your creations from other users.
Midjourney plan prices
These are the US dollar prices listed in its plan comparison, checked in October 2026. Prices change, so check the current price before you subscribe:
| Plan | Monthly price | Billed yearly (per month) | Fast GPU per month | Relax mode | Stealth mode |
|---|---|---|---|---|---|
| Basic | $10 | $8 | 3.3 hours | No | No |
| Standard | $30 | $24 | 15 hours | Unlimited images | No |
| Pro | $60 | $48 | 30 hours | Unlimited images and SD video | Yes |
| Mega | $120 | $96 | 60 hours | Unlimited images and SD video | Yes |
All four plans let you earn free fast time by rating images. Relax mode on the Standard plan helps if you produce client work every week.
What AI images can you create with Midjourney? Use cases
Midjourney allows users to create AI images of almost any kind, from stills to short videos made from an image. It earns its keep in the exploration phase, when you want to see several visual directions before you decide.
- Professional concept art, product photography, portraits and editorial illustration.
- Architecture, interior and product design concepts.
- Mood boards, storyboards and slide backgrounds.
- Five-second videos you can extend 4 seconds at a time, up to 21 seconds.
Its video documentation says clips render in 480p by default and in 720p on the Standard, Pro and Mega plans in fast mode. AI image generation is already common at work. In the EU, 9.5% of enterprises with 10 or more employees used AI to generate pictures, video or sound in 2025, according to Eurostat.
How does Midjourney compare with other AI image generators?
Midjourney is one of several image generators, and each one fits a different way to use AI at work. The table describes how you use each tool, not which one produces better images.
| Tool | How you use it | Typical fit |
|---|---|---|
| Midjourney | Paid subscription on midjourney.com or Discord, with parameters such as –sref and –ar | Exploring visual directions and polished stills |
| ChatGPT | Images created inside a conversation with OpenAI’s assistant | Quick pieces while you work on text |
| Adobe Firefly | Adobe’s family of generative models, used in Adobe’s apps | Teams already working in Adobe tools |
| Stable Diffusion | An open model you can run on your own setup | Technical teams that want full control |
At Founderz we cover several of these tools side by side, because most creative teams end up combining them. Our guide to what Adobe Firefly is explains how Adobe approaches licensing.
Can you use Midjourney images commercially?
Yes, with conditions. The lab’s page on using images and videos commercially says you own what you create, even if you cancel your subscription. You can use it as you like, with two exceptions:
- If you enlarge an image someone else created, it still belongs to its original creator.
- If your company grosses more than $1,000,000 a year, you need a Pro or Mega plan to use the images commercially.
The details sit in section 4 of the terms of service, and the lab notes that it cannot give copyright advice.
What ethical concerns should you consider when using Midjourney?
Four issues come up in real projects: copyright, likeness, bias and disclosure. None of them is solved by a plan, so build the checks into your workflow.
- Copyright: rules for AI-generated art vary by country, so check with a legal advisor before high-stakes use.
- Likeness: avoid prompts that recreate real people, brands or characters.
- Bias: review outputs for stereotypes, since a model reflects the images it learned from.
- Disclosure: tell clients when a visual was made with AI, and keep your instructions as a record.
Key features and limitations of Midjourney AI
The main advantage of this AI software is the aesthetic quality it reaches from simple instructions. Its limits show up when you need total precision, such as long text inside an image or the exact same character across a series.
| Strengths | Limitations |
|---|---|
| Polished aesthetics and a wide range of artistic styles from the first prompt | No free plan |
| Style control with reference images and personalization profiles | Results change between model versions, so lock the version on long projects |
| Editing, canvas extension and video in the same workspace | Open by default: only Pro and Mega plans hide your work with Stealth mode |
| Ownership of the images you generate, under its terms | Companies grossing over $1,000,000 a year need a Pro or Mega plan |
Frequently asked questions
These are common questions people search about Midjourney, answered from its official documentation.
Do you need Discord to use Midjourney?
No. It started on Discord, but today you can create, edit and organize your images on midjourney.com. The official Discord server is still home to the community. You can also work there with the same parameters.
Do you need programming skills for Midjourney?
No. To generate images you write in plain language and add a few parameters at the end, such as –ar for the aspect ratio. The web app guides you through creating, editing and organizing images, so no coding is involved at any step.
How long does it take to generate an image?
It depends on the mode and the model. In fast mode each job uses your fast hours and is ready within seconds in most cases, a little longer in HD. Relax mode does not use fast hours, but your jobs wait in a queue and can take longer.
What is the difference between Midjourney and ChatGPT for images?
Midjourney is dedicated to images and video, with style controls, parameters and model versions. ChatGPT generates images inside a conversation. A practical approach is to draft the prompt in ChatGPT and produce the final image in Midjourney.
Further reading
These links go to the official documentation and to a general reference on the lab’s history.
If you read Spanish and want to move from testing this tool to using it on real projects, Founderz’s training for creatives (in Spanish) shows you how to build it into your workflow. You combine it with Magnific, Adobe Firefly, Runway or Kling and finish with a final project. It is fully online, developed in collaboration with Microsoft and ends with a Founderz and Microsoft certificate.
