instructor

To extract structured data from LLM responses with Pydantic validation, you define a Pydantic model representing your desired schema, and the Skill automatically parses and validates the LLM output against it, ensuring type safety for fields like integers and formatted emails.

When LLM output fails Pydantic validation, the Skill handles validation errors by automatically retrying the LLM call with feedback, prompting the model to correct its output and match the required structured schema.

Yes, you can get structured output from both OpenAI and Anthropic models, as the Skill integrates with various LLM providers to reliably extract structured data using Pydantic schemas across different platforms.

Yes, the Skill supports streaming partial structured results, enabling real-time processing and validation of Pydantic models as the LLM generates the response, rather than waiting for the entire output to complete.

The best way to extract user profiles like name, age, and email from unstructured text is using Pydantic models for structured data extraction, which automatically validates that age is an integer and email is correctly formatted.