Instructor is the most popular Python library for extracting structured data from LLMs using Pydantic validation, making it effortless to get reliably formatted outputs from any language model. Instead of parsing inconsistent LLM text output, Instructor automatically retries with validation feedback until the model produces data matching your exact schema. Supporting all major LLM providers, Instructor handles complex nested structures, lists, enums, and custom validators, making it the go-to tool for production applications that need to extract structured information reliably.