Lift is an open PDF-to-structured-data model designed to move beyond basic text dumping and regex-based extraction for document mining. The project provides a tutorial and benchmark for converting research PDFs into structured JSON using controlled, schema-guided field-level evaluation.