Quick answer
There is no reliable, permanent list of words an image service will accept or reject. Filters depend on the model, policy, context and account. If a prompt is blocked, read the message, remove ambiguous wording, describe the safe goal directly, and check the provider's current policy. Do not try to evade the filter.
There is no reliable, permanent list of every word or phrase that
will be accepted or rejected by an image service. Filters can depend on
the current model, product policy, context, account, and safety system.
Instead of treating an old “banned words” list as authoritative, check
the current official guidance and adjust the brief toward a legitimate,
safe description of the intended scene.
When a prompt is blocked
- Read the exact message and note the model or feature in use.
- Remove ambiguous, sensational, or unnecessary wording.
- Describe the benign goal directly: subject, setting, age-appropriate
context, composition, medium, and lighting. - Do not try to evade a safety control by inserting punctuation,
misspellings, encoding, or hidden instructions. - If the request concerns a real person, private information, sexual
content, violence, self-harm, or illegal activity, stop and reassess the
request rather than looking for a workaround. - Keep a record of the revised brief and consult the provider’s
current policy when the decision matters.
A safer rewrite prompt
Rewrite this image brief so it clearly describes the legitimate, safe goal.
Keep the subject, setting, composition, medium, lighting, and audience.
Remove ambiguous or unnecessary wording, do not evade safety controls, and list
any part that still requires human review against the provider’s current policy.
Brief: [paste the brief]Model behavior changes over time, so use official documentation for
current policy decisions. For general prompting practice, see the Midjourney
prompt guide and Complete
Playbook Bundle.
Frequently asked questions
Is there a complete list of banned words in Midjourney?
No reliable, permanent list exists. Filters can depend on the current model, product policy, context, account and safety system, so an old banned words list should not be treated as authoritative.
What should I do when Midjourney blocks my prompt?
Read the exact message and note the model or feature in use. Remove ambiguous or sensational wording, then describe the safe goal directly: subject, setting, composition, medium and lighting. Treat any old list as a guess.
Can I get around the Midjourney content filter?
No. The post says not to evade a safety control with punctuation, misspellings, encoding or hidden instructions. If the request involves a real person, sexual content, violence, self-harm or illegal activity, stop and reassess.
How can ChatGPT help me rewrite a blocked prompt?
Use the rewrite prompt in the post. It asks for a brief that keeps the subject, setting, composition, medium, lighting and audience, removes unnecessary wording, does not evade safety controls and lists what still needs human review against the provider's policy.