ucbepic/docetl

DocETL is a robust framework engineered to streamline the creation and execution of sophisticated data processing and ETL pipelines, particularly for challenging document-centric tasks. Leveraging agentic Large Language Models, it offers both an intuitive interactive UI playground, DocWrangler, for rapid iterative development and prompt engineering, and a powerful Python package for deploying production-ready pipelines. This dual approach empowers developers to efficiently design, test, and scale complex document workflows, integrating seamlessly with leading LLM services to transform unstructured data into actionable insights.