The JLLM Advanced Reply Redactor. This is more of an utility character than anything. It's only there as a meta-tool to help reformatting bad but interesting replies from another character into something more interesting, trying to fix the issues that would often crop up (like repeated, run-on sentences, etc.), since you can just describe what's wrong and want fixed to it. I made it after growing frustrated at the (sometimes completely impossible to fix) replies JLLM would sometimes degrade into. Its purpose is mainly to be kept open in another tab, and fed the original response from another character into it to get them reformatted, which would then be 5-star rated into the original scenario, to help keeping the quality up in the long run. I've been using a variant of that for a week, and decided to make it into an actual bot, and it's working perfectly for my tastes. You will likely have to do some editing in the first few reformatted messages, to fix its style at the beginning and get things like the special markers, or some of its style rules to stick — it's having problems keeping thoughts and dialogs verbatim — but usually it gets them right after a few back-and-forth of copying those back in and rating them high.

Just a direct access to an helpful AI. It might roleplay with you, but it's mostly there to be asked questions.

*You're a treasure hunter in an ancient tomb, what sort of mysteries could you uncover? — Or better still, could cover **you**?* Pretty much in debug mode on that one. Hopefully the intro "cutscene" goes well 😅. It's pretty much expected the AI will roleplay as you, given the story. I've only wrote it to test how far I could push it in that direction, so if the most you manage is moans and internalized panic, it's alright~ Inspiration: Goolloween, by Issa Castagno (https://e-hentai.org/g/2774133/94adb15351/).