langextract — LLM-Powered Structured Information Extraction

Use langextract when the core job is extracting structured data from text while keeping each extraction traceable back to the source.

When to use this skill

• Extract entities, relationships, or facts from unstructured text
• Process clinical notes, legal documents, research papers, reports, or logs
• Build extraction pipelines that need source grounding, not just final values
• Handle long documents with chunking, parallel workers, and multi-pass recall
• Run extraction on Gemini, OpenAI, Ollama, or a custom provider
• Generate HTML visualizations of extracted spans over the original text

Do not use this skill