AI Agent Glossary
Definitions for the concepts that appear most often when building Python AI agents — from the fundamental patterns to the production infrastructure.
Each page includes a precise definition, how the concept is used in practice, and runnable Python examples using AgentFlow.
Core concepts
| Term | Definition |
|---|---|
| What is an AI agent? | A program that uses an LLM to perceive inputs, reason, call tools, and take actions in a loop to complete multi-step tasks |
| What is a ReAct agent? | An agent that alternates between Reasoning and Acting steps — calling tools, observing results, and reasoning again until it has an answer |
| What is multi-agent orchestration? | Coordinating multiple specialized AI agents so they collaborate on a shared goal, with explicit handoffs and control flow |
| What is a state graph? | A graph-based model for agent workflows where nodes are processing steps and edges define how state moves between them |
| What is agent memory? | The mechanisms by which an AI agent stores and retrieves information across turns, sessions, and restarts |
| What is the Model Context Protocol (MCP)? | An open standard that lets AI agents connect to external tools and data sources through a common interface |
| What is RAG? | Retrieval-Augmented Generation — a pattern where an agent retrieves relevant documents before generating a response |
| What is agent streaming? | Sending AI agent responses token-by-token to a frontend instead of waiting for the full response to complete |
Related
- Compare frameworks — AgentFlow vs LangGraph, CrewAI, AutoGen, Google ADK
- Get started with AgentFlow — build your first agent in Python
- Prebuilt agents — ReactAgent, RAGAgent, SwarmAgent, and more