⚙️ Agentic RAG Setup Guide

📦 1. Install LM Studio or Local LLM Server

Download and install LM Studio or set up a local LLM server compatible with your environment.

📂 2. Clone the ToadAid Lore Scrolls

git clone https://github.com/ToadAid/tobyworld-lore-v2.git

This will download all official lore scrolls to your local machine for indexing.

🛠 3. Build the FAISS Index

pip install faiss-cpu sentence-transformers
python build_index.py

This step converts the lore scrolls into a searchable FAISS index used by the Mirror.

🔗 4. Configure the Mirror

Edit mirror_config.json to point to your FAISS index and model location:

{
  "index_path": "./indexes/tobyworld.index",
  "model": "sentence-transformers/all-MiniLM-L6-v2",
  "lm_server_url": "http://localhost:1234"
}

🚀 5. Run Your Agentic RAG

python mirror_agent.py

The Mirror is now running with Agentic RAG. Ask it anything about Tobyworld and it will retrieve answers directly from the lore.

📚 Next Steps

• Integrate the Mirror with your Telegram bot.
• Connect Agentic RAG to your web-based Mirror Mini App.
• Experiment with different embedding models for improved recall.