Google Document AI vs Unstructured
Pick Google Document AI for bulk OCR on Google Cloud, where Enterprise OCR costs $1.50 per 1,000 pages and drops to $0.60 above 5M pages a month. Pick Unstructured for ingesting mixed file types (PDFs, HTML, Word, PowerPoint, emails, images) into RAG, or when you want an Apache-2.0 library you can self-host. For tables, Document AI scored 64.6% on RD-TableBench, and the open-source Unstructured library scored 0.426 on table cell content in its authors' own benchmark, so test both on your own documents.
Google Document AIgoogle/document-ai | Unstructuredunstructured/unstructured | |
|---|---|---|
| Type | OCR API | OCR API |
| Price | 10 credits / page | 20 credits / page |
| Per 1,000 pages | $10.00 | $20.00 |
| Typical speed | 1.4 s / page | 3.0 s / page |
| Best for | Multilingual scans | Mixed file types into RAG |
| Inputs | PDF, images | PDF, images, DOCX |
| Languages | 200+ languages | Multilingual |
| Max pages per request | 500 | 1,000 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | No | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | ✓ Yes | No |
| Forms | ✓ Yes | No |
| Multilingual | ✓ Yes | ✓ 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 Google Document AI if
- You want plain OCR at volume: $1.50 per 1,000 pages, $0.60 above 5M a month, with the first 1,000 pages free.
- You need RAG chunking from a managed service: Layout Parser includes it at $10 per 1,000 pages.
- You run batch jobs: batch requests take up to 500 pages, though Form Parser batch is capped at 100.
Choose Unstructured if
- Your inputs are mixed: PDFs, HTML, Word, PowerPoint, emails and images go through one library.
- You want to self-host or start free: the library is Apache-2.0, and the hosted platform gives 10,000 free pages, then $0.015 per page.
- You can pick a speed and quality trade-off: the Auto, Fast, High-Res and VLM strategies let you choose, and Fast suits born-digital PDFs.
Key differences
| Google Document AI | Unstructured | |
|---|---|---|
| Licence | Proprietary | Apache-2.0 for the open-source library; proprietary for the hosted platform |
| How you run it | Hosted API on Google Cloud only | Self-host the library or use the hosted API |
| Price | Enterprise OCR $1.50 per 1,000 pages ($0.60 above 5M a month); Layout Parser $10; Form Parser $30 | Hosted: first 10,000 pages free, then $0.015 per page ($15 per 1,000); the open-source library is free |
| Request limits | Online requests up to 15 pages (30 in imageless mode); batch up to 500 pages; Form Parser batch 100 | Not published |
| Output formats | JSON, plain text, bounding boxes, chunks | JSON, HTML, plain text, bounding boxes |
| Table accuracy | 64.6% on Reducto's RD-TableBench, behind Azure and Textract (independent) | Authors' benchmark: table cell content 0.426 for the open-source library versus 0.866 hosted |
| Known failure | Long PDFs in one synchronous call hit the page limit | hi_res tables can cut off at the edges and drop first letters (issue #3262); table OCR ignores language settings (issue #1694); the library sends usage telemetry by default |
Sources
What each one is
Google Document AI
Google Cloud OCR with layout parsing and specialised processors.
Google Document AI detailsUnstructured
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": "google/document-ai","fallbacks": ["unstructured/unstructured"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Google Document AI alternativesUnstructured alternativesBest OCR models in 2026