Comprehensive architecture overview of the NOMOS framework
NOMOS is a sophisticated multi-step agent framework that combines the intelligence of Large Language Models with the reliability and predictability of state machines. This document provides comprehensive architecture diagrams that illustrate the system’s design and component interactions.
Do not get scared by the complexity of the architecture diagrams. They are designed to provide a clear understanding of how NOMOS operates at a low level, and each component plays a crucial role in the overall functionality of the framework.
The NOMOS architecture is designed around several key principles:
Modular Design: Each component has a specific responsibility and can be tested independently
State Machine Reliability: Agent behavior is predictable and auditable through step-based transitions
Flow Organization: Related steps are grouped into flows with shared context and components
Memory Hierarchy: Different levels of memory management for sessions, flows, and context transfer
Tool Integration: Unified interface for custom, external, and package-based tools
API-First Design: RESTful and WebSocket APIs enable diverse client integrations
Production Ready: Built-in persistence, monitoring, and scaling capabilities
This architecture enables NOMOS to bridge the gap between rapid prototyping and production-ready AI agent deployment while maintaining the testability and reliability expected in enterprise software systems.