Docling
Convert document layouts into structured text for an AI retrieval workflow or a source-backed content review.
Community support
Upvotes come from SearchEngineAnswer members. GitHub stars and reader reviews are counted separately.
Claim this tool if you maintain it.
- Best for
- Publishing
- Setup required
- Use Python 3.10 or later in a virtual environment. The chosen document pipeline may download models on its first run.
- Costs to check
- The MIT-licensed library can run locally. Budget for model downloads, disk space and compute. Model licenses and optional service costs need their own review.
- License in catalog
- MIT
GitHub repository
- GitHub stars
- 70,784
- Forks
- 5,383
- Last push
- 2026-11-19
- Primary language
- Python
GitHub snapshot from . Stars count GitHub bookmarks. Reader ratings appear separately below.
How to install Docling
-
Create and activate a Python 3.10 or later virtual environment, then run python -m pip install docling.
python -m pip install docling - Save the example script as convert.py and place report.pdf beside it. Allow the selected pipeline to download its required models before the first conversion.
First tutorial for Docling
-
Choose a PDF you can use that contains two columns, a table and a footnote. Write down one row of table values and the intended paragraph order as a control.
from pathlib import Path from docling.document_converter import DocumentConverter converter = DocumentConverter() result = converter.convert('report.pdf') Path('report.md').write_text( result.document.export_to_markdown(), encoding='utf-8' ) - Run python convert.py and inspect report.md. Check whether the columns follow the intended sequence and the chosen table values retain their units.
- Record any loss of notes or ordering and review the underlying extraction before adding the Markdown to a retrieval system. Repeat with a second document template.
Proposed evaluation using the documented DocumentConverter interface. No extraction benchmark or indexing outcome is claimed.
Where Docling fits
Use Docling to inspect reading order and structured document extraction before indexing a collection of PDFs for retrieval.
- When to choose another tool
- Use a simpler text extractor when a clean HTML or plain-text source already provides the structure and citations you need.
- First evaluation
- Convert a two-column PDF with a known table and compare the exported reading sequence and cell values with the source page.
- How to interpret the result
- An extraction score or well-formed output cannot verify the claims in a document. Check figures, units, footnotes and source links before reuse.
Reviews and discussion
Reader reviews
Be the first to share your experience with this tool.
Discussion
Ask a question or share a setup tip.
Flag an outdated detail
Include the field that needs updating and a source link. Copy your draft, then use the corrections policy to contact the editorial team.
Write a review or join the discussion