MonkeyMind
About
Yo, so I’ve been working on these modules for like 2 months, just some modules to help with AI stuff on Roblox. I thought It’d be cool to share it in case someone else is doing the same kinda thing.
If you’re making anything with LLMs or AI, these might be helpful
Brief Overview
ApePrompt- Makes the prompting a little bit cleaner and easier in the eyes.GorillaMemory- Stores well… memory.- Has In-Session and Persistent memories.
RAG_utan- Simple (and scuffed) Retrieval-Augmented Generation.- It’s basically just an inverted index keyword fuzzy search.
FunctionCaller- Allows the LLM to call functions with the syntax:[FunctionName {params}]
Why I made ts
Prior to this, I was a Functional Extremist, and OOP was a mystery to me. So I made MonkeyMind to force myself to just try it. It’s my first real shot at OOP and I enjoyed it and learned a lot.
There’s probably stuff that could be better, so if you’ve got feedback or run into something weird, I’m down to hear it.
Update Log - 6/11/2025
Changed
- Refactored
RAG_utanto remove OOP usage for a more functional style.
Added
- Support for searching across multiple databases via
RAG_utan.Search(query, {db1, db2}, ...). - New
MaxSnippetsparameter added toRAG_utan.Search, allowing control over the maximum number of results returned.
Example
local Index = RAG_utan.buildIndex(Database)
local Index2 = RAG_utan.buildIndex(Database2)
local Data = {Index, Index2}
local Prediction = RAG_utan.Search("Query Text", Data, 100, 1)
-- QueryText: string
-- Databases: {SearchDatabase}
-- MaxSnippets: number
-- FuzzThreshold: number