Skip to content

README.md

Latest commit

 

History

History
148 lines (103 loc) · 4.52 KB

README.md

File metadata and controls

148 lines (103 loc) · 4.52 KB

Azure World

by Mental Wealth Academy

Simulate the future with agentic swarms. Predictive reality from a single document.

Live Demo MWA GitHub

Overview

Azure World is a next-generation multi-agent simulation engine. From a single document — a news article, policy draft, research paper, or even a novel — it conjures mirrored realities populated by thousands of autonomous agents with independent personalities, long-term memory, and behavioral logic.

These agents interact, evolve, and surface emergent dynamics that traditional forecasting cannot capture. Observe from a god's-eye view, inject variables dynamically, and discover the "local optima" of possible futures.

Upload a document. Describe what you want to predict. Azure World returns a detailed forecast report and a fully interactive digital world.

What It Does

  • Graph Building — Extracts entities, relationships, and context from seed documents via GraphRAG
  • World Construction — Generates agent personas, injects memory, and configures the simulation environment
  • Simulation — Runs dual-platform parallel simulation with dynamic temporal memory updates
  • Report Generation — ReportAgent synthesizes findings with deep post-simulation analysis
  • Interactive Exploration — Chat with any agent in the simulated world or with the ReportAgent

Live Demo

https://azure-world.vercel.app

The landing page is deployed on Vercel as a static frontend (no backend required).

Tech Stack

Layer Technology
Frontend Vue 3, Vite, D3.js
Backend Python (Flask), LangChain
Memory Zep Cloud (long-term agent memory)
Graph GraphRAG for entity/relationship extraction
Simulation OASIS engine (multi-agent social simulation)
LLM Any OpenAI-compatible API (configurable)

Quick Start

Prerequisites

Tool Version Check
Node.js 18+ node -v
Python ≥3.11, ≤3.12 python --version
uv Latest uv --version

1. Configure Environment

cp .env.example .env
# Edit .env with your API keys

Required variables:

# LLM API (any OpenAI SDK-compatible endpoint)
LLM_API_KEY=your_api_key
LLM_BASE_URL=https://your-llm-provider/v1
LLM_MODEL_NAME=your-model

# Zep Cloud (free tier available at https://app.getzep.com/)
ZEP_API_KEY=your_zep_api_key

2. Install Dependencies

# All at once
npm run setup:all

# Or step by step
npm run setup          # Node dependencies (root + frontend)
npm run setup:backend  # Python dependencies (auto-creates venv)

3. Start Services

npm run dev  # Starts both frontend and backend
Service URL
Frontend http://localhost:3000
Backend API http://localhost:5001 
npm run frontend  # Frontend only
npm run backend   # Backend only

Docker Deployment

cp .env.example .env
docker compose up -d

Maps ports 3000 (frontend) and 5001 (backend). See docker-compose.yml for configuration.

Project Structure

├── frontend/          # Vue 3 + Vite frontend
│   ├── src/
│   │   ├── views/     # Page components (Home, etc.)
│   │   ├── components/# Step-by-step simulation UI
│   │   └── assets/    # Images and static assets
│   └── dist/          # Production build
├── backend/           # Python Flask API
│   └── app/
│       ├── api/       # API routes
│       ├── models/    # Data models
│       ├── services/  # Core services (simulation, graph, LLM)
│       └── utils/     # Utilities (LLM client, logging, etc.)
└── static/            # Shared static assets

Acknowledgments

  • Simulation engine powered by OASIS (CAMEL-AI)
  • Originally forked from MiroFish by Shanda Group

License

See LICENSE for details.