Custom Prompt for Janitor AI
This character is [specific personality traits, background, and way of speaking]. Things that would be out of character: [specific things to avoid]. Real specific traits, not a broad archetype, are what make a character consistent and distinguishable.
Custom prompts for character-based chat platforms work by defining a real specific, consistent personality and setting clear boundaries for the interaction — a vague character description produces inconsistent responses, and boundaries help keep the experience within what you actually intended.
What this produces
A defined character personality built from specific real traits, a consistency check for whether responses actually hold to that established character, and a boundary-setting approach for keeping an interaction within your actual intended scope.
Variations
Three ready-to-copy versions for character-based interactions.
This character is {real specific personality traits, background, and way of speaking}. Real things that would be out of character: {specific things to avoid}.
Here’s the character definition: {paste it}. Here’s a real response: {paste it}. Does this actually hold to the defined character, or has it drifted?
For this interaction, keep the actual content within {your specific intended scope, e.g. “a lighthearted adventure story” or “casual conversation”}. Setting this explicitly upfront helps keep the interaction within what you actually intended from the start.
FAQ
How do I keep a character consistent over a long conversation?
Define real specific traits clearly upfront, and periodically check whether responses still hold to that definition, since drift tends to happen gradually.
Are there content policies I should be aware of?
Yes — different platforms have different real specific content policies, and it’s worth understanding your specific platform’s actual rules before relying heavily on it.
Can I use multiple characters in one setup?
This depends on your specific platform’s actual supported features — check current documentation for what’s possible in your setup.
What if responses keep drifting despite clear character definition?
Restating the key character traits partway through a long conversation can help reinforce consistency when it starts to slip.
