Docling vs Unstructured
Pick Docling for RAG ingestion of born-digital PDFs and Office files when you want a self-hosted MIT library that exports lossless JSON with layout and table structure. Pick Unstructured when you also want a hosted API, with the first 10,000 pages free, or when you ingest mixed file types through connectors. Watch tables: the open-source Unstructured library scores 0.426 on table cell content against 0.866 hosted, and Docling's scan accuracy depends on the OCR engine you plug in.
Doclingibm/docling | Unstructuredunstructured/unstructured | |
|---|---|---|
| Type | Open source | OCR API |
| Price | 2 credits / page | 20 credits / page |
| Per 1,000 pages | $2.00 | $20.00 |
| Typical speed | 2.2 s / page | 3.0 s / page |
| Best for | Structured export for RAG | Mixed file types into RAG |
| Inputs | PDF, images, DOCX | PDF, images, DOCX |
| Languages | Latin scripts, CJK | Multilingual |
| Max pages per request | 500 | 1,000 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | No | ✓ Yes |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose Docling if
- It parses PDF, DOCX, PPTX, XLSX, HTML, EPUB, images, emails and XBRL through one API.
- Its lossless JSON export keeps layout, reading order and table structure for downstream tools.
- It is MIT licensed and runs fully local, on CPU or GPU, with a 258M GraniteDocling VLM option and pluggable OCR engines (EasyOCR, Tesseract, RapidOCR).
Choose Unstructured if
- It offers a hosted API with the first 10,000 pages free, then $0.015 per page, so you do not have to run anything yourself.
- Its Auto, Fast, High-Res and VLM strategies let you trade speed for quality per job, and the library comes with many connectors.
- It returns typed elements with bounding boxes, and the fast strategy suits born-digital PDFs.
Key differences
| Docling | Unstructured | |
|---|---|---|
| How you run it | Local Python library only; you pay for compute | Self-hosted open-source library or hosted API |
| Licence | MIT (individual models have their own licences) | Apache-2.0 library; proprietary hosted platform |
| Price | Free, self-hosted | Library free; hosted is 10,000 pages free, then $0.015 per page; Business is custom |
| Output formats | Markdown, HTML, JSON, DocTags, plain text | JSON, HTML, plain text, bounding boxes |
| Table accuracy | Not published | Table cell content 0.426 (open-source) vs 0.866 (hosted), vendor's own benchmark |
| Known failure | 2.143.1 drops words and numbers with full-page Tesseract OCR (#5254); reading order can merge neighbouring columns (#5132) | hi_res tables cut off at edges and drop first letters (#3262); table OCR ignores language settings (#1694) |
| Scanned pages | Accuracy depends on the OCR engine you plug in | Extra Tesseract pass on tables adds processing time (#2580) |
Sources
- Docling GitHub README
- Supported formats
- Issue #5254 OCR regression
- Unstructured pricing
- Unstructured GitHub README
- Unstructured issue #3262
What each one is
Unstructured
Document ETL for RAG. Partitions many file types into typed elements; open-source library too.
Unstructured detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "ibm/docling","fallbacks": ["unstructured/unstructured"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Docling alternativesUnstructured alternativesBest OCR models in 2026