An open-source collection of practitioner guides covering the full stack of agentic AI — from foundations and design patterns to safety, evaluation, and production deployment.
Core concepts and mental models for understanding agentic AI systems — perception loops, reasoning strategies, and an honest map of the frameworks available today. All four modules are now published with full prose, code examples, and diagrams.
Read Chapter 01 →“An agent is not just an LLM with tools — it is a system that perceives, decides, and acts in a loop, with the capacity to recover from failure.”
— agentic-ai / agents-introArchitectural blueprints for building agents that are reliable, composable, and maintainable — from simple single-agent loops to complex multi-agent orchestration. All four modules are now published with full prose, code examples, and annotated diagrams.
Read Chapter 02 →How agents remember, retrieve, and act on the world — covering every memory type from in-context buffers to vector databases, and tool use from function calls to browser automation. All four modules are now published with full prose, code examples, and diagrams.
Read Chapter 03 →“Memory is what separates a stateless chatbot from a true agent. Without it, every conversation starts from zero.”
— agentic-ai / memoryTrust, guardrails, and measuring whether your agent actually works — covering prompt injection defences, evaluation harnesses, tracing infrastructure, and responsible AI governance. All four modules are now published with full prose, code examples, and diagrams.
Read Chapter 04 →“An agent that cannot be evaluated cannot be trusted. Evals are not optional — they are the foundation of production readiness.”
— agentic-ai / evals