Concepts

The mental model behind LangExtract: the extraction pipeline, how grounding ties results back to the source, and how a model_id selects a provider.

📄️How extraction works

This page builds the mental model behind lx.extract. You don't need it to run

Grounding ties every extraction back to the exact span of source text it came

LangExtract does not hardcode a single model. The model_id you pass to