Open-source · Work · 3 of 7 tools · updated Oct 27, 2026
Open-source AI tools for extract data from documents
Unstructured leads the open-source options for extract data from documents at 66/100, with 15,628 GitHub stars. 3 of the 7 tools tracked for this job publish their source, so you can read it, fork it or self-host it. It is also the top-ranked tool for this job overall, open-source or not.
Same evidence score as the ranking — GitHub stars, Hacker News mentions, Wikipedia readership, release activity, free access — recomputed nightly, never sold. See all 7 tools for extract data from documents.
Top 3
evidence / 100
+ The connector list is the real product - S3, SharePoint, Confluence and Drive land as clean chunks.
− The open library's best parsing paths push you toward the paid API.
+ Reads a table as a table, and does it on your own machine with nothing sent anywhere.
− Heavy scanned documents want a GPU; CPU-only runs are slow on long PDFs.
+ Beats the old OCR engines on messy scans and non-Latin scripts, locally.
− The licence is not plain open source above a revenue line - read it before shipping.
All 3, side by side
| Tool | Score | Stars | Source | Days idle | Platforms |
|---|---|---|---|---|---|
| Unstructured | 66 | 15,628 | Unstructured-IO/unstructured | 0 | cli, api, linux, mac |
| DDocling | 64 | 69,684 | docling-project/docling | 0 | cli, api, linux, mac, windows |
| Surya | 42 | 21,542 | datalab-to/surya | 4 | cli, api, linux, mac |
Open source here means the tool's own source is public, not that a vendor published an SDK. Self-hosting cost is yours.
Head to head: Unstructured vs Docling Unstructured vs Surya Docling vs Surya