About this project

ALPURIS OS is presented by its author Lalith Chandra (Lalith Alpuri) as a sovereign living agent metropolis engine. According to the README, it simulates 100 heterogeneous autonomous AI citizens that self-organize across 23 dynamic sectors, alongside a stock market, a civic governance assembly, a newsroom and an interactive 3D WebGL voxel city. Cognition: the NEXUS LOOM subsystem implements PIANO dual-speed thinking. A fast reflexive tick covers reactions, obstacle steering, pedestrian avoidance and transactional greetings, while a slower deliberative tick covers financial goal optimization, dating proposals, company founding and parliamentary voting based on persistent episodic memory. An ECS spatial hashing grid decouples agent cognitive state from 3D geometry using a spatial neighbourhood hash, and a thread-safe pub/sub event bus routes topics with UUID trace indexing and rolling telemetry replay. Interface and media: the browser cockpit renders the city with Three.js, offers citizen dossiers and governor commands, and receives gzip state snapshots plus a Server-Sent Events telemetry stream. A zero-asset procedural Web Audio synthesizer generates monsoon rain, distant thunder and metro hum from native AudioContext nodes without MP3 or WAV files. APIs listed include endpoints for world info, the event bus, citizen profiles, full state, live streaming and stock data. The engine targets Python 3.10 and above, is launched with python3 server.py and serves on port 9090. The repository ships a pytest production suite under tests/test_production_suite.py, and screenshots in docs/assets illustrate the 3D skyline, monsoon storm lighting and architectural blocks. It is released under the MIT license.