Docling Project

Open-source document AI infrastructure

Docling Project is the open-source project organization around Docling, a document processing toolkit used to convert PDFs, Office files, HTML, images, audio, video, and other formats into structured outputs for generative AI systems. The GitHub repository describes Docling as a way to get documents ready for gen AI, with local execution, OCR, table structure, layout understanding, Markdown and JSON export, and integrations with frameworks such as LangChain and LlamaIndex. The project originated with IBM Research Zurich and is hosted under the LF AI & Data Foundation. LF AI & Data notes that IBM donated Docling to the foundation as an incubation-stage project in March 2025 and that it graduated to a higher project tier in August 2026. For builders, Docling Project matters because document ingestion is one of the most common bottlenecks in RAG, agent workflows, and model fine-tuning pipelines. The organization page gives OpenTools a canonical parent for Docling-related tools, resources, and future MCP integrations.