Skip to content

The Agent Loop (ReAct) Explained

An agent is a model in a loop: think, act, observe, repeat. Interleaving reasoning with tool calls grounds every step in what the last step actually returned — and every loop needs a brake.

Animated diagrams · 2 flows

Compare all 2

Sort by any column to find the right fit. Click a name to open its details.

Every Loop Needs a BrakeLoop state → stop decisionYes — every turn checksOverhead is one check per turnYour harness, not the modelLow — three boolean gatesProduction agents with real spend
The ReAct LoopThought → action → observation → thoughtYes — multi-step runsOne round trip per tool callModel + toolsMedium — needs guardrailsTasks where the next step depends on the last result