Docling Fundamentals

Docling performs structure-aware extraction on PDFs, DOCX, and HTML, transforming unstructured documents into structured, metadata-rich representations. It preserves layout, sections, and tables to enable knowledge bases, RAG pipelines, and automated metadata capture across formats.

Convert HTML and scanned documents to Markdown or JSON by applying Docling's structure-aware parsing. This multi-format document-processing approach captures page numbers and section titles while exporting provenance-tracking metadata for downstream processing.

Yes, Docling preserves document provenance and section titles during extraction. It performs structure-aware parsing to capture page numbers, layout, and section metadata, ensuring the structured output retains original document context for AI workflows.

The best way to capture metadata from multi-format enterprise documents is using Docling to transform unstructured inputs into structured representations. It extracts section titles and page numbers from PDFs and HTML, exporting to DoclingDocument for automated metadata capture.

Yes, you can use Docling output directly with downstream knowledge base tools. It exports structured documents to Markdown, JSON, or DoclingDocument formats, enabling seamless integration with RAG pipelines and enterprise document processing systems.

You need to install Docling to start parsing documents, but no additional dependencies are required. Simply install Docling, create a DocumentConverter, and convert your sample PDF or DOCX to see structured output for your AI workflows.