AI Glossary Term

What does “ReAct” mean?

Definition

A prompting pattern that interleaves reasoning steps with tool-use actions to solve tasks more reliably.

How ReAct works

Diagram for ReAct, a repeating loop: Thought: what do I need?; then Action: call a tool; then Observation: read the result; then Thought again, then answer; then back to the start.Thought: what do I need?Action: call a toolObservation: read the resultThought again, then answerrepeat

Example

Asked who directed a film and where they were born, the model writes a thought, searches for the film, reads the result, then searches the director and answers using what it found.

Why it matters

Interleaving reasoning with actions ties thinking to real information. It made early tool-using agents more reliable on question answering.

Common questions

What does ReAct stand for?

Reasoning and Acting. It comes from a 2022 paper by researchers at Princeton and Google.

How is it different from chain-of-thought?

Chain-of-thought reasons internally. ReAct mixes reasoning with calls to outside tools.

Is it still used?

The pattern underlies many agent frameworks, though newer models often have tool use built in.

Learn more in our free guides

Test yourself

See also