Poker Simulator

An AI-powered Texas Hold'em simulator where 3–5 LLM agents with distinct personalities (Calculator, Shark, Gambler, Maniac, Rock) play poker against each other. Supports three AI backends: local Ollama models, the Gemini API, or Gemini via Chrome automation.

Full Documentation

---

Prerequisites

• Python 3.11+
• Node.js 18+
• Ollama (for Ollama mode) — with qwen3.5:latest or llama3.2:latest pulled
• A Gemini API key (for API mode)

---

Setup

1. Clone and configure environment

git clone <repo-url>
cd poker
cp backend/.env.example backend/.env
# Edit backend/.env and add your GEMINI_API_KEY if using API mode

2. Install backend dependencies

cd backend
pip install -r requirements.txt
# If using browser mode, also run:
playwright install chromium

3. Install frontend dependencies

cd frontend
npm install

4. Start the backend

cd backend
uvicorn main:app --reload

5. Start the frontend

cd frontend
npm run dev

Open http://localhost:5173 in your browser.

---

AI Modes

Mode	Description	Requirements
Ollama	Runs a local model via Ollama	OLLAMA_URL in .env, model pulled
API	Calls Gemini API directly	GEMINI_API_KEY in .env
Browser	Automates Gemini via Chrome	Click "Init Browser" in settings

---

RAG / Book Ingestion (optional)

To give the AI agents poker strategy knowledge, ingest books into the vector database:

1. Place .txt poker strategy files in backend/books/
2. Run: cd backend && python scripts/ingestBooks.py

The agents will automatically retrieve relevant excerpts when making decisions.