AI-powered digital twin simulation for smart cities, factories, energy grids, and logistics. Free, open-source, research-grade — runs entirely on your machine.
From agent modeling to 3D visualization, from IoT data ingestion to distributed scaling — all in one platform.
Four agent types with customizable behaviors.
Grid and graph worlds with 8 zone types.
Deep learning with graceful fallbacks.
Interactive Three.js world with full controls.
Connect real-world sensors to your simulation.
Full twin framework with GIS and plugins.
Scale across multiple nodes.
Reproducible, exportable, configurable.
Docker Compose gets you running instantly.
Four ready-to-run simulations covering smart cities, factories, energy grids, and emergency response.
Urban traffic simulation with vehicles and pedestrians across residential, commercial, and industrial zones.
Production line optimization with machines, workers, and energy constraints.
Multi-source energy grid with solar plants and varying demand patterns.
System resilience testing under power outages and machine breakdowns.
S(t+1) = F(S(t), A(t), E(t)) — every step is deterministic, extensible, and research-grade.
React Dashboard
Three.js 3D View
Canvas 2D
Metrics Charts
FastAPI REST
WebSocket
CLI Tool
Auto-docs
Engine S(t+1)=F()
Event Bus
State Manager
Snapshot System
4 Agent Types
Grid + Graph
8 Zone Types
Resources
LSTM Predictor
Anomaly Detector
LP Optimizer
RL Agents
MQTT Source
File Source
API Source
Alert Manager
GIS / GeoJSON
Plugins
Marketplace
Connector
Multi-Node
Partitioning
Load Balancer
gRPC
From core engine to full digital twin — every planned feature is implemented and production-ready.
Core engine with S(t+1)=F(S(t),A(t),E(t)), 4 agent types, grid/graph worlds, AI prediction + LP optimization, 4 scenarios, REST + WebSocket API, React 2D dashboard, Docker Compose.
PyTorch LSTM predictor with NumPy fallback, autoencoder anomaly detection, Gymnasium RL environment, PPO/Q-learning agents, multi-agent AI (Planner, Predictor, Optimizer), adaptive feedback loops.
React Three Fiber 3D world, orbit/pan/zoom, 3D zones + agent objects with glow, day/night cycle, seamless 2D↔3D switching.
MQTT topic subscription, CSV/JSON file ingestion, REST API polling, synthetic simulator with noise/drift injection, ring buffer, alert manager.
Distributed engine, simulation nodes with heartbeat, spatial partitioning, load balancing with migration, gRPC protocol, message serialization.
Digital twin core (live/replay/hybrid), GIS with GeoJSON and geofencing, plugin system with hot-reload, marketplace client, REST/WebSocket connector, API key auth, rate limiting.
No Docker? Use Python only:
Join the open-source digital twin community. Build, contribute, and push the boundaries of simulation.