Skip to main content
Docling is available as a Python package and works on macOS, Linux, and Windows. Both x86_64 and arm64 architectures are supported.

​
Requirements

Python 3.9 support was dropped in Docling version 2.70.0. Python 3.10 or higher is required.
Docling supports the following Python versions:
  • Python 3.10
  • Python 3.11
  • Python 3.12
  • Python 3.13
  • Python 3.14

​
System compatibility

  • macOS (x86_64 and Apple Silicon arm64)
  • Linux (x86_64 and arm64)
  • Windows (x86_64)

​
Basic installation

​
Using pip

The simplest way to install Docling is via pip: 
pip install docling
This installs the core package with all essential dependencies for document conversion.

​
Using uv

uv is a fast Python package installer. Install Docling with: 
uv pip install docling

​
Using conda

If you use conda for package management: 
conda install -c conda-forge docling
Note: Conda package availability may lag behind PyPI releases. Check conda-forge for the latest available version.

​
Optional dependencies

Docling offers several optional feature sets that you can install based on your needs.

​
OCR engines

For enhanced OCR capabilities, install one or more OCR backends:
EasyOCR supports 80+ languages with deep learning models:
pip install docling[easyocr]

​
Visual Language Models (VLM)

Use Visual Language Models for enhanced document understanding: 
pip install docling[vlm]
This enables models like GraniteDocling with MLX acceleration on Apple Silicon.

​
Audio processing

For audio and video file processing with Automatic Speech Recognition: 
pip install docling[asr]
Includes Whisper models with MLX acceleration on Apple Silicon.

​
XBRL financial documents

To process XBRL (eXtensible Business Reporting Language) documents: 
pip install docling[xbrl]

​
GPU acceleration

For GPU-accelerated inference (Linux with CUDA): 
pip install docling[onnxruntime]
This installs onnxruntime-gpu for faster model inference on NVIDIA GPUs.

​
Combined installation

Install multiple optional features together: 
pip install docling[easyocr,vlm,asr]

​
Verify installation

After installation, verify that Docling is working correctly:
1

Check the CLI

Verify the command-line interface is available:
docling --version
2

Test Python import

Open a Python interpreter and import Docling:
from docling.document_converter import DocumentConverter
print("Docling imported successfully!")
3

Run a test conversion

Convert a simple document to verify everything works:
from docling.document_converter import DocumentConverter

converter = DocumentConverter()
result = converter.convert("https://arxiv.org/pdf/2408.09869")
print("Conversion successful!")
print(result.document.export_to_markdown()[:200])

​
Troubleshooting

​
Installation issues

If you see errors about Python version:
ERROR: Package 'docling' requires a different Python: 3.9.0 not in '>=3.10,<4.0'
Upgrade to Python 3.10 or higher. Check your version with:
python --version
If you encounter dependency conflicts, try installing in a fresh virtual environment:
python -m venv docling-env
source docling-env/bin/activate  # On Windows: docling-env\Scripts\activate
pip install docling
Some optional features require system libraries:For Tesseract on Ubuntu/Debian:
sudo apt-get update
sudo apt-get install tesseract-ocr libtesseract-dev
For Tesseract on macOS:
brew install tesseract
On Apple Silicon Macs (M1, M2, M3), ensure you’re using native arm64 Python:
python -c "import platform; print(platform.machine())"
Should output arm64. If it shows x86_64, you’re running Rosetta Python.

​
Network issues

Docling downloads AI models from Hugging Face on first use. If you’re behind a firewall:
  1. Set proxy environment variables: 
    export HTTP_PROXY=http://proxy.example.com:8080
export HTTPS_PROXY=http://proxy.example.com:8080
  2. Pre-download models to a local cache and set HF_HOME: 
    export HF_HOME=/path/to/local/cache
  3. Use air-gapped environments by copying the cache directory to offline systems.

​
Development installation

To contribute to Docling or use the latest development version: 
git clone https://github.com/docling-project/docling.git
cd docling
pip install -e .
For development with all tools: 
pip install -e ".[dev,docs,examples]"
See CONTRIBUTING.md for detailed development setup.

​
Next steps

Quick start

Convert your first document with a complete example

Basic conversion

Learn the fundamentals of document conversion

CLI usage

Use Docling from the command line

Configuration

Configure pipelines and customize behavior

Build docs developers (and LLMs) love

Get started for freeTalk to us