Docling is an open-source document processing library that parses a wide range of formats — including PDF, DOCX, PPTX, EPUB, HTML, audio, video, and more — into a unified representation ready for generative AI use. It features advanced PDF understanding with layout analysis, table extraction, formula recognition, and OCR support for scanned documents. Docling integrates natively with popular AI frameworks such as LangChain, LlamaIndex, Crew AI, and Haystack, and can connect to agents via an MCP server. It supports local execution for sensitive or air-gapped environments and can be deployed as a standalone API service.