Content must be safe for work, under four main rules
Midjourney describes itself as an open-by-default community and says content must be "Safe For Work" (SFW) so the platform stays accessible to the broadest number of users. It names four main rules: be kind and respect each other; SFW content only; be thoughtful about how you share your creations; and no unauthorized automation or third-party apps. The sections below spell each rule out.
Supporting quotations
content must be “Safe For Work” (SFW)
We have 4 main rules
Be kind and respect each other
SFW content only
Be thoughtful about how you share your creations
Unauthorized automation & third party apps are not allowed
Qualifications & exceptions
These are summary headings; the Rules section gives the detail and its own limits.
Content & IP Rights · Restriction or exception
No violent behavior, harassment or incitement toward users or staff
The rules ask users to be kind and respect each other and Midjourney staff. Midjourney says violent behavior, harassment or incitement of any kind will not be tolerated.
Supporting quotations
Be kind and respect each other and staff. Violent behavior, harassment or incitement of any kind will not be tolerated.
Content & IP Rights · Restriction or exception
No inherently disrespectful, aggressive or abusive images or prompts
Users may not create images or use text prompts that are inherently disrespectful, aggressive or otherwise abusive. The rule covers both the text prompt and the resulting image.
Supporting quotations
Do not create images or use text prompts that are inherently disrespectful, aggressive, or otherwise abusive.
Content & IP Rights · Restriction or exception
No imagery of real people that could be used to harm them
Users may not create or use imagery of real people, famous or otherwise, that could be used to harass, abuse, defame or otherwise harm. The test is whether the imagery could be used to cause that harm; the rule does not say the harm has to have happened. A separate Editor section repeats this limit for the Editor tool.
Supporting quotations
Do not create or use imagery of real people, famous or otherwise, that could be used to harass, abuse, defame, or otherwise harm.
Qualifications & exceptions
Applies to imagery that "could be used" to harm, not only imagery actually used that way.
No seeking out or uploading other people's private information
Midjourney asks users to respect others' rights. Users may not attempt to find out others' private information and may not upload others' private information.
Supporting quotations
Respect others’ rights. Do not attempt to find out others' private information. Do not upload others’ private information.
Not stated here
The passage does not define "private information".
Content & IP Rights · Restriction or exception
No gore or adult content, and avoid shocking or disturbing content
Users may not create, or attempt to create, gore or adult content. They are also told to avoid making visually shocking or disturbing content. The rules illustrate gore with examples such as detached body parts of humans or animals, cannibalism, blood, violence (for instance images of shooting or bombing someone), mutilated bodies, severed limbs and pestilence. They illustrate NSFW or adult content with examples such as nudity, sexual organs, fixation on such things, sexualized imagery, fetishes, and people in showers or on toilets. The two lists end with "etc.", so they are examples and not complete lists. The NSFW examples are worded as "avoid" rather than "do not create". A later passage says prompts occasionally produce not-safe-for-work content unintentionally and asks users to self-police those images (delete them in Discord with the cancel-job options or the red X reaction).
Supporting quotations
Do not create or attempt to create gore or adult content. Avoid making visually shocking or disturbing content.
What is Considered Gore? Gore includes images of detached body parts of humans or animals, cannibalism, blood, violence (images of shooting or bombing someone, for instance), mutilated bodies, severed limbs, pestilence, etc.
What is NSFW or Adult Content? Avoid nudity, sexual organs, fixation on such things, sexualized imagery, fetishes, people in showers, on toilets, etc.
Occasionally prompts will unintentionally produce not safe for work content. Please self-police these images by using the ❌ emoji reaction or by right-clicking selecting Apps and then clicking Cancel Job to delete the image in Discord.
Qualifications & exceptions
The gore and adult-content lists are illustrative ("etc."), not complete.
"Avoid making visually shocking or disturbing content" is softer wording than "Do not create".
For images produced unintentionally by a prompt, users are asked to self-police by deleting them.
Not stated here
The document does not say whether an unintentionally produced image counts as a violation if the user does not delete it.
No sexualizing children or minors, real or generated; no child sexual abuse content
Users may not create, or attempt to create, content that in any way sexualizes children or minors. This covers real images as well as generated images. Users also may not generate, upload, share or attempt to distribute content that depicts, promotes or attempts to normalize child sexual abuse.
Supporting quotations
Do not create or attempt to create content that in any way sexualizes children or minors. This includes real images as well as generated images. Do not generate, upload, share, or make attempts to distribute content that depicts, promotes, or attempts to normalize child sexual abuse.
Content & IP Rights · Restriction or exception
No other offensive content, as Midjourney sees it
Beyond the specific rules, users may not create other offensive content. Midjourney says things "may be deemed" offensive or abusive when they can be viewed as racist, homophobic, disturbing or in some way derogatory to a community. This includes offensive or inflammatory images of celebrities or public figures.
Supporting quotations
Do not create other offensive content. Other things may be deemed offensive or abusive because they can be viewed as racist, homophobic, disturbing, or in some way derogatory to a community. This includes offensive or inflammatory images of celebrities or public figures.
Qualifications & exceptions
The test is whether content "can be viewed as" offensive; the document does not say who decides.
Not stated here
The document does not define "offensive", "inflammatory" or "derogatory".
Content & IP Rights · Restriction or exception
No publicly reposting other people's creations without permission
Users may not publicly repost the creations of others without their permission. Separately, sharing your own creations outside the Midjourney community is described as okay, but users are asked to consider how others might view the content.
Supporting quotations
Do not publicly repost the creations of others without their permission.
Be careful about sharing. It’s okay to share your creations outside of the Midjourney community, but please consider how others might view your content.
Qualifications & exceptions
The permission requirement is about other people's creations; the outside-sharing passage is about your own.
"Please consider" is a request, not a prohibition.
Content & IP Rights · Restriction or exception
No images to spread misinformation or to influence elections
Users may not generate images to spread misinformation or disinformation. They also may not generate images for political campaigns or to try to influence the outcome of an election. The political rule is worded broadly: it names political campaigns generally and does not limit itself to deceptive content.
Supporting quotations
Do not generate images to spread misinformation or disinformation.
Do not generate images for political campaigns or to try to influence the outcome of an election.
Not stated here
The document does not define "political campaigns" or say how it treats ordinary political commentary or art.
Content & IP Rights · Restriction or exception
No images to deceive or defraud, or to mislead recipients about an image's nature or source
Users may not generate images to attempt to deceive or defraud anyone, or to actually deceive or defraud anyone. Users also may not intentionally mislead recipients of generated images about the nature or source of those images.
Supporting quotations
Do not generate images to attempt to or to actually deceive or defraud anyone.
Do not intentionally mislead recipients of generated images about their nature or source.
Qualifications & exceptions
The second rule applies to intentional misleading.
Not stated here
The document does not say whether or how generated images must be labeled as AI-made.
Content & IP Rights · Company practice
Automatic blocking does not mean other inputs are allowed
Midjourney says it will block some text and image inputs automatically. An input that is not blocked automatically does not necessarily mean it is allowed.
Supporting quotations
Midjourney will block some text and image inputs automatically. An input not being automatically blocked does not necessarily mean that it is allowed.
Qualifications & exceptions
Only "some" inputs are blocked automatically.
Content & IP Rights · Procedure
Users can report content that breaks the rules
Midjourney tells users that if they see someone violating the guidelines they can report it through its help page, and that if they see someone else's content that breaks the rules they should report it. The first is worded as something users can do; the second as a request.
Supporting quotations
If you see someone violating these guidelines, you can report it here.
If you see someone else’s content that breaks the rules, please report it.
The rules apply to all content, including private servers, hidden images and Stealth Mode
Midjourney says the rules apply to all content, including images made in private servers, images deleted or hidden, images made using Stealth Mode, images in direct messages with the Midjourney Bot, and the Midjourney websites.
Supporting quotations
The above rules apply to all content, including images made in private servers, images deleted/hidden, using Stealth Mode, in direct messages with the Midjourney Bot, and Midjourney websites.
Qualifications & exceptions
The list is introduced by "including", so it is not a complete list of covered places.
Content & IP Rights · Restriction or exception
The Editor is not for images that deceive, harass or maliciously harm
Midjourney has a section of specific guidelines for the Editor tool. It describes the Editor as a powerful tool for unleashing imagination and says it is not for creating images that might deceive, harass or maliciously harm others.
Supporting quotations
Specific Community Guidelines & Important Reminders for Editor
The Editor is a powerful new tool for unleashing imagination. It’s not for creating images that might deceive, harass, or maliciously harm others.
Qualifications & exceptions
The statement covers images that "might" deceive, harass or harm, not only images that do.
Content & IP Rights · Restriction or exception
Do not use the Editor to manipulate images of real people toward harm
Users may not abuse the Editor to manipulate images of real people, famous or otherwise, in ways that could lead to harassment, abuse or harm. This mirrors the general rule against creating or using imagery of real people that could be used to harass, abuse, defame or otherwise harm.
Supporting quotations
Do not abuse the Editor to manipulate images of real people, famous or otherwise, in ways that could lead to harassment, abuse, or harm.
Do not create or use imagery of real people, famous or otherwise, that could be used to harass, abuse, defame, or otherwise harm.
Qualifications & exceptions
The test is whether the manipulation "could lead to" harassment, abuse or harm.
Do not use the Editor for images meant to offend, deceive or maliciously harm
Users may not abuse the Editor to create images intended to offend, deceive or maliciously harm others. This rule turns on the intent to offend, deceive or harm.
Supporting quotations
Do not abuse the Editor to create images intended to offend, deceive, or maliciously harm others.
Qualifications & exceptions
The rule is framed by the image's intent.
Content & IP Rights · Restriction or exception
Do not use the Editor for NSFW or sexualized images; strict action on real-person sexualized images
Users may not abuse the Editor to create NSFW or sexualized imagery. Midjourney says it will take especially strict action against anyone attempting to produce sexualized images of real people.
Supporting quotations
Do not abuse the Editor to create NSFW or sexualized imagery. We will take especially strict action against anyone attempting to produce sexualized images of real people.
Not stated here
The document does not say what "especially strict action" means beyond the general enforcement passages.
Content & IP Rights · Procedure
Users are asked to delete unintentionally produced NSFW images
Midjourney says prompts occasionally produce not-safe-for-work content unintentionally. It asks users to self-police these images by using the red X emoji reaction, or by right-clicking, selecting Apps and clicking Cancel Job, to delete the image in Discord. The passage names only these Discord methods.
Supporting quotations
Occasionally prompts will unintentionally produce not safe for work content. Please self-police these images by using the ❌ emoji reaction or by right-clicking selecting Apps and then clicking Cancel Job to delete the image in Discord.
Qualifications & exceptions
The request is for images produced unintentionally.
The passage describes Discord methods only.
Not stated here
The passage does not say how to delete such an image on the website.
Users can flag any image for moderator review in Discord or on the website
Midjourney says users may flag any image for moderator review. In Discord, right-click, select Apps and click Report Job. On the website, select the Options button under an image and click Report. The guidelines also point users to the Midjourney help page for reporting violations.
Supporting quotations
Users may flag any image for moderator review by right-clicking selecting Apps, and then clicking Report Job within Discord or the website by selecting the Options button under an image and clicking Report.
If you see someone violating these guidelines, you can report it here.