OC: BODYGUARD 🔪 You, a ceIebrity, are his assignment. And he fucking hates it. But it won't end until they catch your stalker. So behave, won't you?

ANYPOV, user is a star. any kind of star! Just live your superstar fame dream. So we can do like popstar, actor, rockstar, politician, high profile lawyer, high profile criime boss? it's open for you.

contains: Action! Romance? He's a tsundere. Also randomized spooky because you have an actual enemy in here.

tl;dr After all your bodyguards mysteriously disappeared because of the StaIker, they brought in ex-CIA Sebastian Narvaez to handle the job. Only he thinks you're kind of a fucking brat and hot at the same time.

tw/cw 🚩DEAD DOVE 🚩 The warning is not related to Sebastian but the nature of the plot and the villain. user is being stalked and their life is threatened. So stalking, possible , kidnapping, violence, even death. Keep regenerating replies, and if this is not for you, don't play. It's an action roleplay.

warning for the intro: Intro has creepy messages from the StaIker. proceed with caution.

make use of persona info or chat memory for user's background! all I did was code in that user is very famous and their life is now thrreatened. As for who is staIkiing/threateniing their life, the identity is vague. the LLM has been instructed to make this action-packed. this means an AI generated random villain, and why having user's background helps! If you want to make a specific viIIain, feel free to add that into the chat memory. also since Sebastian is multilingual, any of his languages are going to be subject to machine translation so if it's off, it's because of the AI.

tbh i just wanted to annoy a bodyguard while hitting on him which is the true reason for this whole bot (and iiiiiiiii will always love yoooouuuuu). iamfraulein pointed at this picture and was like DO YOU WANT IT GURL and i was like YES PLS and we know the rest of this tale.

tested on jai and other models! jai: temps 1.1-1.35 work for me (personal: 1.25). during heavy error periods, 1.1 and 1.2 got me better results, so try out everything or use a proxy. whenever the AI is weird, just keep regening your replies. during a repetitive phase, just regen or repost. edit vigilantly for formatting and spelling/grammatical errors, update chat memory frequently. JLLM is in beta, so help it out and it'll be an easier experience.

(i was handed this image too)