SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
Project Links
Meta
Author: Christoph Auer, Michele Dolfi, Maxim Lysak, Nikos Livathinos, Ahmed Nassar, Panos Vagenas, Peter Staar
Requires Python: <4.0,>=3.10
Classifiers
Operating System
- MacOS :: MacOS X
- POSIX :: Linux
- Microsoft :: Windows
Development Status
- 5 - Production/Stable
Intended Audience
- Developers
- Science/Research
Topic
- Scientific/Engineering :: Artificial Intelligence
Programming Language
- Python :: 3
- Python :: 3.10
- Python :: 3.11
- Python :: 3.12
- Python :: 3.13
- Python :: 3.14
Docling
Docling simplifies document processing, parsing diverse formats โ including advanced PDF understanding โ and providing seamless integrations with the gen AI ecosystem.
Features
- ๐๏ธ Parsing of multiple document formats incl. PDF, DOCX, PPTX, XLSX, HTML, WAV, MP3, WebVTT, images (PNG, TIFF, JPEG, ...), LaTeX, plain text, and more
- ๐ Advanced PDF understanding incl. page layout, reading order, table structure, code, formulas, image classification, and more
- ๐งฌ Unified, expressive DoclingDocument representation format
- โช๏ธ Various export formats and options, including Markdown, HTML, WebVTT, DocTags and lossless JSON
- ๐ Support of several application-specifc XML schemas incl. USPTO patents, JATS articles, and XBRL financial reports.
- ๐ Local execution capabilities for sensitive data and air-gapped environments
- ๐ค Plug-and-play integrations incl. LangChain, LlamaIndex, Crew AI & Haystack for agentic AI
- ๐ Extensive OCR support for scanned PDFs and images
- ๐ Support of several Visual Language Models (GraniteDocling)
- ๐๏ธ Audio support with Automatic Speech Recognition (ASR) models
- ๐ Connect to any agent using the MCP server
- ๐ป Simple and convenient CLI
What's new
- ๐ค Structured information extraction [๐งช beta]
- ๐ New layout model (Heron) by default, for faster PDF parsing
- ๐ MCP server for agentic applications
- ๐ผ Parsing of XBRL (eXtensible Business Reporting Language) documents for financial reports
- ๐ฌ Parsing of WebVTT (Web Video Text Tracks) files and export to WebVTT format
- ๐ฌ Parsing of LaTeX files
- ๐ Parsing of plain-text files (
.txt,.text) and Markdown supersets (.qmd,.Rmd)
Coming soon
- ๐ Metadata extraction, including title, authors, references & language
- ๐ Chart understanding (Barchart, Piechart, LinePlot, etc)
- ๐ Complex chemistry understanding (Molecular structures)
Installation
To use Docling, simply install docling from your package manager, e.g. pip:
pip install docling
Note: Python 3.9 support was dropped in docling version 2.70.0. Please use Python 3.10 or higher.
Works on macOS, Linux and Windows environments. Both x86_64 and arm64 architectures.
More detailed installation instructions are available in the docs.
Getting started
To convert individual documents with python, use convert(), for example:
from docling.document_converter import DocumentConverter
source = "https://arxiv.org/pdf/2408.09869" # document per local path or URL
converter = DocumentConverter()
result = converter.convert(source)
print(result.document.export_to_markdown()) # output: "## Docling Technical Report[...]"
More advanced usage options are available in the docs.
CLI
Docling has a built-in CLI to run conversions.
docling https://arxiv.org/pdf/2206.01062
You can also use ๐ฅGraniteDocling and other VLMs via Docling CLI:
docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062
This will use MLX acceleration on supported Apple Silicon hardware.
Read more here
Documentation
Check out Docling's documentation, for details on installation, usage, concepts, recipes, extensions, and more.
Examples
Go hands-on with our examples, demonstrating how to address different application use cases with Docling.
Integrations
To further accelerate your AI application development, check out Docling's native integrations with popular frameworks and tools.
Get help and support
Please feel free to connect with us using the discussion section.
Technical report
For more details on Docling's inner workings, check out the Docling Technical Report.
Contributing
Please read Contributing to Docling for details.
References
If you use Docling in your projects, please consider citing the following:
@techreport{Docling,
author = {Deep Search Team},
month = {8},
title = {Docling Technical Report},
url = {https://arxiv.org/abs/2408.09869},
eprint = {2408.09869},
doi = {10.48550/arXiv.2408.09869},
version = {1.0.0},
year = {2024}
}
License
The Docling codebase is under MIT license. For individual model usage, please refer to the model licenses found in the original packages.
LF AI & Data
Docling is hosted as a project in the LF AI & Data Foundation.
IBM โค๏ธ Open Source AI
The project was started by the AI for knowledge team at IBM Research Zurich.
2.84.0
Apr 01, 2026
2.83.0
Mar 31, 2026
2.82.0
Mar 25, 2026
2.81.0
Mar 20, 2026
2.80.0
Mar 14, 2026
2.79.0
Mar 12, 2026
2.78.0
Mar 10, 2026
2.77.0
Mar 06, 2026
2.76.0
Mar 02, 2026
2.75.0
Feb 24, 2026
2.74.0
Feb 17, 2026
2.73.1
Feb 13, 2026
2.73.0
Feb 11, 2026
2.72.0
Feb 03, 2026
2.71.0
Jan 30, 2026
2.70.0
Jan 23, 2026
2.69.1
Jan 21, 2026
2.69.0
Jan 20, 2026
2.68.0
Jan 13, 2026
2.67.0
Jan 09, 2026
2.66.0
Dec 24, 2025
2.65.0
Dec 15, 2025
2.64.1
Dec 09, 2025
2.64.0
Dec 02, 2025
2.63.0
Nov 20, 2025
2.62.0
Nov 17, 2025
2.61.2
Nov 10, 2025
2.61.1
Nov 06, 2025
2.61.0
Nov 06, 2025
2.60.1
Nov 04, 2025
2.60.0
Oct 31, 2025
2.59.0
Oct 30, 2025
2.58.0
Oct 22, 2025
2.57.0
Oct 15, 2025
2.56.1
Oct 13, 2025
2.56.0
Oct 13, 2025
2.55.1
Oct 03, 2025
2.55.0
Sep 30, 2025
2.54.0
Sep 22, 2025
2.53.0
Sep 17, 2025
2.52.0
Sep 11, 2025
2.51.0
Sep 05, 2025
2.50.0
Sep 03, 2025
2.49.0
Sep 01, 2025
2.48.0
Aug 26, 2025
2.47.1
Aug 23, 2025
2.47.0
Aug 22, 2025
2.46.0
Aug 20, 2025
2.45.0
Aug 18, 2025
2.44.0
Aug 12, 2025
2.43.0
Jul 28, 2025
2.42.2
Jul 24, 2025
2.42.1
Jul 22, 2025
2.42.0
Jul 18, 2025
2.41.0
Jul 10, 2025
2.40.0
Jul 04, 2025
2.39.0
Jun 27, 2025
2.38.1
Jun 25, 2025
2.38.0
Jun 23, 2025
2.37.0
Jun 16, 2025
2.36.1
Jun 04, 2025
2.36.0
Jun 03, 2025
2.35.0
Jun 02, 2025
2.34.0
May 22, 2025
2.33.0
May 20, 2025
2.32.0
May 14, 2025
2.31.2
May 13, 2025
2.31.1
May 12, 2025
2.31.0
Apr 25, 2025
2.30.0
Apr 14, 2025
2.29.0
Apr 10, 2025
2.28.4
Mar 29, 2025
2.28.3
Mar 28, 2025
2.28.2
Mar 26, 2025
2.28.1
Mar 25, 2025
2.28.0
Mar 19, 2025
2.27.0
Mar 18, 2025
2.26.0
Mar 11, 2025
2.25.2
Mar 05, 2025
2.25.1
Mar 03, 2025
2.25.0
Feb 26, 2025
2.24.0
Feb 20, 2025
2.23.1
Feb 20, 2025
2.23.0
Feb 17, 2025
2.22.0
Feb 14, 2025
2.21.0
Feb 10, 2025
2.20.0
Feb 07, 2025
2.19.0
Feb 07, 2025
2.18.0
Feb 03, 2025
2.17.0
Jan 28, 2025
2.16.0
Jan 24, 2025
2.15.1
Jan 10, 2025
2.15.0
Jan 08, 2025
2.14.0
Dec 18, 2024
2.13.0
Dec 17, 2024
2.12.0
Dec 13, 2024
2.11.0
Dec 12, 2024
2.10.0
Dec 09, 2024
2.9.0
Dec 09, 2024
2.8.3
Dec 03, 2024
2.8.2
Dec 03, 2024
2.8.1
Nov 29, 2024
2.8.0
Nov 27, 2024
2.7.1
Nov 26, 2024
2.7.0
Nov 20, 2024
2.6.0
Nov 19, 2024
2.5.2
Nov 13, 2024
2.5.1
Nov 12, 2024
2.5.0
Nov 12, 2024
2.4.2
Nov 08, 2024
2.4.1
Nov 08, 2024
2.4.0
Nov 04, 2024
2.3.1
Oct 30, 2024
2.3.0
Oct 30, 2024
2.2.1
Oct 28, 2024
2.2.0
Oct 23, 2024
2.1.0
Oct 18, 2024
2.0.0
Oct 16, 2024
1.20.0
Oct 11, 2024
1.19.1
Oct 11, 2024
1.19.0
Oct 08, 2024
1.18.0
Oct 03, 2024
1.17.0
Oct 03, 2024
1.16.1
Sep 27, 2024
1.16.0
Sep 27, 2024
1.15.0
Sep 24, 2024
1.14.0
Sep 24, 2024
1.13.1
Sep 23, 2024
1.13.0
Sep 18, 2024
1.12.2
Sep 17, 2024
1.12.1
Sep 16, 2024
1.12.0
Sep 13, 2024
1.11.0
Sep 10, 2024
1.10.0
Sep 10, 2024
1.9.0
Sep 03, 2024
1.8.5
Aug 30, 2024
1.8.4
Aug 30, 2024
1.8.3
Aug 28, 2024
1.8.2
Aug 27, 2024
1.8.1
Aug 26, 2024
1.8.0
Aug 23, 2024
1.7.1
Aug 23, 2024
1.7.0
Aug 22, 2024
1.6.3
Aug 22, 2024
1.6.2
Aug 22, 2024
1.6.1
Aug 21, 2024
1.6.0
Aug 20, 2024
1.5.0
Aug 20, 2024
1.4.0
Aug 14, 2024
1.3.0
Aug 12, 2024
1.2.1
Aug 07, 2024
1.2.0
Aug 07, 2024
1.1.2
Jul 31, 2024
1.1.1
Jul 30, 2024
1.1.0
Jul 26, 2024
1.0.2
Jul 24, 2024
1.0.1
Jul 24, 2024
1.0.0
Jul 18, 2024
0.4.0
Jul 17, 2024
0.3.1
Jul 17, 2024
0.3.0
Jul 17, 2024
0.2.0
Jul 16, 2024
0.1.2
Jul 15, 2024
0.1.1
Jul 15, 2024
0.1.0
Jul 15, 2024
Wheel compatibility matrix
Files in release
Extras:
Dependencies:
pydantic
(<3.0.0,>=2.0.0)
docling-core[chunking]
(<3.0.0,>=2.70.0)
docling-parse
(<6.0.0,>=5.3.2)
docling-ibm-models
(<4,>=3.13.0)
torch
(<3.0.0,>=2.2.2)
torchvision
(<1,>=0)
filetype
(<2.0.0,>=1.2.0)
pypdfium2
(!=4.30.1,<6.0.0,>=4.30.0)
pydantic-settings
(<3.0.0,>=2.3.0)
huggingface_hub
(<2,>=0.23)
requests
(<3.0.0,>=2.32.2)
ocrmac
(<2.0.0,>=1.0.0)
rapidocr
(<4.0.0,>=3.3)
certifi
(>=2024.7.4)
rtree
(<2.0.0,>=1.3.0)
typer
(<0.22.0,>=0.12.5)
python-docx
(<2.0.0,>=1.1.2)
python-pptx
(<2.0.0,>=1.0.2)
beautifulsoup4
(<5.0.0,>=4.12.3)
pandas
(<4.0.0,>=2.1.4)
marko
(<3.0.0,>=2.1.2)
openpyxl
(<4.0.0,>=3.1.5)
lxml
(<7.0.0,>=4.0.0)
pillow
(<13.0.0,>=10.0.0)
tqdm
(<5.0.0,>=4.65.0)
pluggy
(<2.0.0,>=1.0.0)
pylatexenc
(<3.0,>=2.10)
scipy
(<2.0.0,>=1.6.0)
accelerate
(<2,>=1.0.0)
polyfactory
(>=2.22.2)
defusedxml
(<0.8.0,>=0.7.1)