A modular, cross-model universal prompt.
This is a build-your-own-prompt project, separated into sections you can mix and match. Unlike other customizable prompts, the sections here address common biases and issues rather than stylistic choices. Keep in mind a jack-of-all-trades like this will never be as optimal as a carefully engineered dedicated prompt, but it's a good start for any model recent enough not to have one yet.
Sections that can be toggled are marked with a [?]
, and they start with a comment explaining why you might need it.
Some example setups:
Writing for {{user}}
, Anti-cooking
, Anti-positivity
, Unblock NSFW
, Anti-horny
Writing for {{user}}
, Anti-repetition
, Anti-cooking
, Anti-positivity
, Unblock NSFW
, Anti-horny
Anti-cooking
, Anti-positivity
, Unblock NSFW
, Story start
, Reinforcement
, NSFW reinforcement
chatgpt-4o-latest
.)