An agent doesn't answer in one shot — it loops. It thinks, picks a tool, observes the result, and repeats until the goal is met. This is the ReAct loop, in motion.
ThinkThe LLM decides what to do next.
ActIt calls a tool — search, a DB, code.
ObserveIt reads the tool's output.
Repeat → AnswerLoops until the goal is met.
A single LLM call just predicts text. An agent wraps it in a loop with tools and memory — so it can take actions, check results, and self-correct toward a goal. Multi-agent systems run several of these in concert. You build agent systems with LangGraph in FORGE.