docling is a Python library that parses diverse document formats, including advanced PDF understanding, for generative AI applications. It provides a unified document representation format with integrations into popular generative AI frameworks.
Project overview
The combination of advanced PDF layout analysis, multi-format parsing, and integration with generative AI frameworks like LangChain and LlamaIndex makes docling a practical tool for document processing pipelines.
Project type
RAG · Document Processing · Data Processing
Use cases
Documents & Office · Knowledge Q&A
Deployment
Refer to project documentation
License
MIT
Best for
Developers and AI engineers building retrieval-augmented generation pipelines who need to convert diverse document formats into a unified representation.
Data teams processing documents that contain complex layouts, tables, formulas, code, or images for generative AI ingestion.
Key capabilities
Analyzes page layout, reading order, table structure, code, formulas, and image classification within PDF documents.
Provides extensive OCR support, Visual Language Models including GraniteDocling, and audio support with Automatic Speech Recognition (ASR) models.
Provides a CLI for document conversion and a Python library API for programmatic use.
Metadata extraction including title, authors, references, and language is a planned capability.
Getting started
Run pip install docling, then run docling https://arxiv.org/pdf/2206.01062 to convert a document from a URL.
GitHub project description: Get your documents ready for gen AI
README: Docling simplifies document processing by parsing diverse formats — including advanced PDF understanding — and providing seamless integrations with the generative AI ecosystem.
README: - 🔒 Local execution capabilities for sensitive data and air-gapped environments - 🤖 Plug-and-play [integrations][integrations] incl. LangChain, LlamaIndex, Crew AI & Haystack for…