Structured Output
A mode of LLM generation in which the model is constrained to produce output conforming to a predefined schema—typically JSON—enabling reliable downstream parsing without post-processing heuristics.
Structured output solves a fundamental integration problem: LLMs naturally produce free-form text, but downstream systems need machine-readable data. Early approaches used regex post-processing to extract structured data from text, which was brittle. Constrained decoding (grammar-based sampling, logit masking) enforces schema conformance at the token level, guaranteeing valid output.
OpenAI introduced structured outputs in 2024; Anthropic supports it via tool-use schema enforcement and, with Claude Sonnet 4.6+, via direct JSON mode. The practical effect is that agents can reliably emit structured action plans, classification decisions, or database records without the scaffolding needing a fallback parsing strategy.
Structured output is most valuable for: agent-to-agent communication (typed payloads reduce parsing errors), data extraction from documents (financial reports, contracts), and any workflow where downstream code consumes LLM output programmatically. The cost is slight: constrained decoding can marginally reduce output quality by preventing the model from using intermediate reasoning tokens.
## Related Terms
- [[glossary/json-mode|JSON Mode]] — the model output mode that constrains generation to valid JSON - [[glossary/function-calling|Function calling]] — the mechanism by which the model emits structured JSON to invoke an external function - [[glossary/tool-use|Tool use]] — the protocol Claude follows when invoking external tools from an agent loop - [[glossary/sampling-parameters|Sampling Parameters]] — the temperature, top-k, and top-p settings that control output randomness - [[courses/claude-tool-use-from-zero|Course: Claude Tool Use from Zero]] — hands-on practice with the concepts covered in this entry