Introducing DocLang, a Markup Language for AI
IBM, Wednesday, August 12th, 2026
IBM Research introduces DocLang, an AI-native document markup standard following its Docling parser.
IBM Research introduces DocLang, a markup language designed specifically for AI consumption of documents. The post is built around an interview with Peter Staar, the lead researcher behind IBM's open-source document parser Docling, on why generative AI needs its own document standard.
The argument is that business-critical data is often locked in outdated or proprietary formats, which limits what enterprises can extract from charts, customer files and annual reports.
Docling has become IBM's most successful open-source project ever, with 32 million downloads and 64,000 GitHub stars, and has been integrated into LangChain, Red Hat OpenShift and IBM's Granite model family. DocLang is presented as the next step: an AI-native document representation standard.