Glossary
What is Midjourney?
In short
Midjourney is a service from Midjourney, Inc. that creates images and short videos from a text description.
Models and products · Midjourney
MMOOZZGG.com editorsChecked: 1 min read
For a child
Imagine an artist you tell: “Draw a cat astronaut on the Moon”, and a minute later the artist shows you four different drawings. You pick one and ask to add stars or make it brighter. There is a program that draws from words just like that. But the pictures don’t always turn out the way you meant, and you must not take other people’s drawings without asking.
In plain words
It is a service that draws from a description: you write a phrase about what you want to see, and it returns several image options. You work with it on its website and in Discord. You can improve a finished picture in the editor or bring it to life: since June 2025 the service makes short clips from an image. For example, a marketer describes the packaging of a new tea in autumn colors and quickly gets a dozen sketches to discuss with a designer. By default, the images you create are visible to other members of the community.
For business
Under the terms of service as of 27 May 2026, a company or its employee with revenue above 1 million dollars a year gets the rights to images only with a Pro or Mega subscription. Everything you create is public by default and open to remixing; a private mode (Stealth) comes with the Pro and Mega plans. The user also gives the company a perpetual license to their prompts and results. The terms have no list of supported countries, and Russia is not named there (checked 16 October 2026). The service has no version for your own servers. So images of unannounced products are not uploaded there, and rights to pictures for advertising are discussed with a lawyer. How to build image generation into marketing work without leaks is covered on the page about generative AI.
For a specialist
Midjourney is a text-to-image system. The company does not disclose its architecture or training data, so you can judge how it works only from the documentation and announcements. Version V7, announced on 4 April 2025, was the first with model personalization turned on by default. It added Draft Mode, which, according to the company’s announcement, costs half as much and renders 10 times faster at the cost of quality; you re-render a draft you like in full quality. The V1 video model (18 June 2025) works as image-to-video: one job gives four clips of 5 seconds each, and each can be extended by about 4 seconds, up to four times. Under the terms, users must be at least 13 years old.
Example
A coffee shop orders a seasonal menu from a designer, who first makes several design sketches in the service to agree on the style quickly. The designer finishes the final illustrations by hand and checks that they don’t look like someone else’s work.
Related terms
Questions and answers
Sources
Not sure whether your company needs this?
AI solutions map