Working with AI is always very finicky, but from my knowledge there aren’t many limitations on what you can write with SudoWrite. (Looking at what plug-ins are popular, many users are able to write explicit erotic content with it)
From what I’ve noticed, pretty much all AI language models tend to more evasive or vague when they “feel” that the topics described are not appropriate in this context. Always make sure the AI is properly encouraged though tone or other context that these elements you don’t want to get sanitized are appropriate to use.
This effect is particularly prevalent when your idea goes through AI multiple times: When you for example create a story by using Braindump → Generate Synopsis → Generate Outline → Generate Scenes → Generate Draft, your original concept will become more smoothed out with every single step in the process. The shorter your process, the better AI will retain your sharp edges or potentially “inappropriate” instructions.
There are two strategies I like to use against the AI trying to sanitize: - Check and edit AI output (synopsis, outlines, character cards, scenes, …) after each step and make sure that everything you want to happen is as explicitly stated as possible and doesn’t contradict anything else the AI has access to. - With every step when using AI, clearly define the tone and language you want. Try not to beat around the bush and give the AI examples. In case you want to write explicit content, using explicit or even profane language when interacting with the AI can help a lot.
Log in to comment and vote
Comments1
Vaturna
Dec 23, 2025
Working with AI is always very finicky, but from my knowledge there aren’t many limitations on what you can write with SudoWrite. (Looking at what plug-ins are popular, many users are able to write explicit erotic content with it)
From what I’ve noticed, pretty much all AI language models tend to more evasive or vague when they “feel” that the topics described are not appropriate in this context. Always make sure the AI is properly encouraged though tone or other context that these elements you don’t want to get sanitized are appropriate to use.
This effect is particularly prevalent when your idea goes through AI multiple times: When you for example create a story by using Braindump → Generate Synopsis → Generate Outline → Generate Scenes → Generate Draft, your original concept will become more smoothed out with every single step in the process. The shorter your process, the better AI will retain your sharp edges or potentially “inappropriate” instructions.
There are two strategies I like to use against the AI trying to sanitize:
- Check and edit AI output (synopsis, outlines, character cards, scenes, …) after each step and make sure that everything you want to happen is as explicitly stated as possible and doesn’t contradict anything else the AI has access to.
- With every step when using AI, clearly define the tone and language you want. Try not to beat around the bush and give the AI examples. In case you want to write explicit content, using explicit or even profane language when interacting with the AI can help a lot.