instructor - Extract and validate structured LLM outputs
Extracts typed, validated structured data from LLM responses, retries validation failures, parses JSON, and streams partial results.
Tags
Updated: 2026-10-02Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Extract structured data
- Validate Pydantic models
- Retry failed validations
- Parse typed JSON
- Stream partial objects
- Support multiple providers
Inputs
- Pydantic response schema
- Source text
- Prompt messages
- LLM model identifier
- Provider API credentials
Outputs
- Validated structured objects
- Parsed JSON data
- Partial streamed objects
- Streamed iterable items
- Validation error feedback
Requirements
- Python environment
- instructor package
- pydantic package
- LLM provider SDK
- Provider API access
- Structured-output model support