— Concepts
Structured Output
Forcing an LLM to return JSON / typed data that matches a schema you define.
Also known as: JSON mode · Schema-constrained generation · Typed output
What is Structured Output?
Structured output is the technique of constraining an LLM to return data in a precise format — usually JSON conforming to a schema you specify. Without it, you get prose that needs parsing; with it, you get an object you can plug straight into your code. Modern LLMs (Claude, GPT-4o, Gemini) support structured output natively via response_format or tool-use schemas. It is essential for any production AI feature: classification, extraction, agent state machines, form filling, lead scoring.
— Related
Terms connected to Structured Output
Concepts
Function Calling
An LLM feature that lets the model invoke your code/APIs in a structured way.
Open →Concepts
Tool Use
An AI's ability to call external tools (APIs, code, search) instead of just generating text.
Open →Concepts
AI Agent
An AI system that decides its own next action and takes multi-step actions autonomously.
Open →Techniques
Prompt Engineering
The craft of designing inputs to LLMs to get reliable, high-quality outputs.
Open →From definitions to deployed projects.
Knowing what a term means is step one. ONROL's AI Generalist track gets you shipping projects that use it.