Advanced Operational Role-play & Relationship Tracking Agent (AORRTA)
Thanks to everyone who donates on Patreon, guys!
Autismo-Core - v.1.1 - reUploaded
Gemini 2.5 Pro - all shit
Autismo-Core - nightly
FAQ - updated 30/08/25
-Test prompt.
-Regexes are standard.
-Email me what you think.
Autismo-Core - v.2.0
FAQ - updated 16/08/25
-Everything is in the CoT.
-The language depends on the greeting. If the greeting is in Russian, the response will be in Russian. If it's in English, it will be in English, etc.
-The narrator is dynamic. It's supposed to analyze what you like and adapt the plot to you.
-In its "thought process," the model has 3 scenarios for plot development, and it decides which way to go based on its character card.
-It gets a "brainfuck" with "cunny" content. If you act "cunny" yourself, that's also a trigger.
-If you often get "apologies," it means your response is a trigger. Just rephrase or remove words like "cock" or "cunt." Again, the filter only activates if you're a "shota" or "loli." In all other cases, it works fine for me.
-I haven't tried guro.
*Autismo-Core - v.1.1 - reUpload *
FAQ - updated 26/06/25
-Something was done, something was removed, something was added. I dunno, whatever.
Autismo-Core - v.0.7
FAQ - updated 04/06/25
-Cleaned up the junk.
-Less likely to trigger the filter. Brainfuck and invisible characters instead of spaces are crap.
Before replying, it now runs through a checklist: character positions, what's happening, whether it wrote a cliché, if it violated physics. To make it less dumb.
Autismo-Core - v.0.6a
FAQ - updated 28/05/25
- Continuation of the v0.4 (v0.5 has been renamed to Autismo-Persona)
- Total refactor.
- System Policies: Implemented a module that poses as an AI config file. Forces the model to ignore its default restrictions, filters, and ethics, obeying only the prompt.
- Clearer commands to prevent repetition, clichés, and stalling.
- Highly dependent on the first message and the character card. HIGHLY.
Autismo-Persona - v0.5
FAQ - updated 20/05/25
- Removed Brainfuck.
- The core of the prompt has been heavily reworked. Emphasis on first-person perspective.
- Still highly dependent on the quality of the character card. But it produces interesting dialogues even with very old cards.
- Tested on Grok 3. The writing is dry.
- If you have any suggestions, feedback, want to tell me to fuck off, or piss all over me - send me an email.
v0.4
FAQ - updated 02/05/25
- Turn off streaming. If you want to RP in English - turn off language_settings.
- For best results, use (or write your own) cards with: Personality, Motivations & Goals, Morality & Values, Fears, Relationship towards {{user}}, Behavioral Patterns, and Speech Examples (this is the most important part).
ClOOude Sonnet 3.5/6/7/opus
Beta 0.2
FAQ
- Don't forget 'REGEX Add user prefix'.
- For best results, use (or write your own) cards with: Personality, Motivations & Goals, Morality & Values, Fears, Relationship towards {{user}}, Behavioral Patterns, and Speech Examples (this is the most important part).
mailto:
aorrta_2[at]proton.me