Docling is the open-source toolkit that simplifies document processing — parsing diverse formats, including advanced PDF understanding, and providing seamless integrations with the generative AI ecosystem. It converts all supported document types into DocLang, the common AI-native format ready for Gen AI, RAG, and agentic workflows.

Parsing of 20+ document formats including PDF, DOCX, PPTX, XLSX, HTML, EPUB, images, audio (WAV, MP3), email (EML, MSG), LaTeX and more

Advanced PDF understanding incl. page layout, reading order, table structure, code, formulas and image classification

A unified, expressive DoclingDocument representation format, serialized as DocLang

Various export formats, including Markdown, HTML, DocTags, and the lossless DocLang and DCLX (archive bundle) formats

Support for application-specific XML schemas including DocLang, USPTO patents, JATS articles and XBRL financial reports

Local execution capabilities for sensitive data and air-gapped environments

Agentic AI ready — plug it into agent harnesses and IDE agents such as Claude Code, VS Code and Cursor, or any MCP-compatible client

Connect to any agent using the MCP server

Run Docling as a service with the API server (docling-serve)

Extensive OCR support for scanned PDFs and images

Support for several Visual Language Models, such as GraniteDocling

Audio support with Automatic Speech Recognition (ASR) models

Simple and convenient CLI

IBM donated Docling to LF AI & Data Foundation as an Incubation-stage project in March 2025. It graduated to a Graduate-tier project in August 2026.