Docling

Package

Open-source Python library that parses PDFs, Office files and images

Price: Free, open source
Access: None, runs locally

About

MIT-licensed Python library and CLI that turns PDFs, Office files, HTML, images and audio into Markdown or lossless JSON, keeping layout, reading order and tables, with OCR for scans. It runs locally, even air-gapped, and needs Python 3.10 or later.

What you can do with it

Convert PDFs, DOCX, PPTX and XLSX files into clean Markdown or JSON for RAG
Extract tables, reading order and page layout from complex PDF reports
OCR scanned PDFs and images on your own machine without sending data out

Get started

Install with pip install docling (Python 3.10+)
Pass a local path or URL to DocumentConverter

Example

pip install docling
python - <<'EOF'
from docling.document_converter import DocumentConverter

result = DocumentConverter().convert("https://arxiv.org/pdf/2408.09869")
print(result.document.export_to_markdown())
EOF

Details

Hosting: Self-hosted, Runs locally
Available in: Worldwide
Official SDKs: Python
MCP server: Local