toy project. it's used for learning AutoGPT, LangChain and other things related to LLMs.

My goal is to build a customized AI assistant that can help me understand codebase, organize informations such as article or books, and increase the efficiency of my daily work.

I want to use the architecture from AutoGPT.

Basically, we have Agent, Planning, Resources, Abilities and Memory. I've removed Plugins and Workspace since i don't know how to utilize them right now.

TODO(sheep): more architecture details

python3 -m venv .venv
source .venv/bin/activate
cd arcane_intelligence
poetry install
python main.py

• Implement LLM provider, chat with LLM first
• [] Test LLAMA Index
  • [] Implement llama index based on MoonshotAI
• [] Test LangChain
• [] Implement Planning
• [] Implement Abilities
• [] Combine above components and implement simple agent
• [] Introduce memory subsystem