Installation¶

Overview¶

Docling Graph is available on PyPI. Install with pip for the recommended experience, or clone the repository and use uv for development.

What You'll Install¶

Core Package: Docling Graph with VLM support
Optional Features: LLM providers (local and/or remote) via LiteLLM (included by default)
GPU Support (optional): PyTorch with CUDA for local inference
API Keys (optional): For remote LLM providers

Quick Start¶

Install from PyPI (Recommended)¶

pip install docling-graph

This installs:

✅ Docling (document conversion)
✅ VLM backend (NuExtract models)
✅ Core graph functionality
✅ LiteLLM (for LLM providers; no extra install needed)

Run the CLI with:

docling-graph --version
docling-graph --help

Install from Source (Development)¶

To contribute or use the latest development version:

git clone https://github.com/docling-project/docling-graph
cd docling-graph
uv sync

Use uv run docling-graph when running the CLI from a source checkout.

System Requirements¶

Minimum Requirements¶

Python: 3.10, 3.11, or 3.12
RAM: 8 GB minimum
Disk: 5 GB free space
OS: Linux, macOS, or Windows (with WSL recommended)

Recommended for Local Inference¶

GPU: NVIDIA GPU with 8+ GB VRAM
CUDA: 11.8 or 12.1
RAM: 16 GB or more
Disk: 20 GB free space (for models)

For VLM Only¶

GPU: NVIDIA GPU with 4+ GB VRAM (for NuExtract-2B)
GPU: NVIDIA GPU with 8+ GB VRAM (for NuExtract-8B)

For Remote API Only¶

No GPU required
Internet connection required
API keys required

Verification¶

Check Installation¶

# Check version (use docling-graph if installed via pip; uv run docling-graph if from source)
docling-graph --version

# Test CLI
docling-graph --help

Expected output:

Docling Graph v1.2.0
Usage: docling-graph [OPTIONS] COMMAND [ARGS]...

Test Import¶

python -c "import docling_graph; print(docling_graph.__version__)"

Expected output:

v1.2.0

Next Steps¶

After installation, you need to:

Set Up Requirements - Verify system requirements
Configure GPU (optional) - Set up CUDA for local inference
Set Up API Keys (optional) - Configure remote providers
Define Schema - Create your first Pydantic template

Common Issues¶

🐛 `uv` not found (source install only)¶

If you install from source, you need uv. Install it with:

# Linux/macOS
curl -LsSf https://astral.sh/uv/install.sh | sh

# Windows (PowerShell)
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

# Or with pip
pip install uv

🐛 Python version mismatch (source install)¶

When using uv from source, specify Python version if needed:

uv python install 3.10
uv sync

🐛 Import errors after installation¶

Solution: If you installed from source with uv, run scripts and the CLI via uv run:

uv run python script.py
uv run docling-graph --help

If you installed with pip, use python and docling-graph directly.

🐛 GPU not detected¶

Solution: See GPU Setup Guide

Performance Notes¶

New in v1.2.0: Significant CLI performance improvements:

Init command: 75-85% faster with intelligent dependency caching
First run: ~1-1.5s (checks dependencies)
Subsequent runs: ~0.5-1s (uses cache)
Dependency validation: 90-95% faster (2-3s → 0.1-0.2s)
Lazy loading: Configuration constants loaded on-demand

Development Installation¶

For contributing to the project:

# Clone repository
git clone https://github.com/docling-project/docling-graph
cd docling-graph

# Install with development dependencies
uv sync --all-extras --dev

# Install pre-commit hooks
uv run pre-commit install

# Run tests
uv run pytest

Updating¶

If you installed from PyPI:

pip install -U docling-graph

If you installed from source:

git pull origin main
uv sync

Uninstalling¶