Google Document AI vs dots.ocr
Pick dots.ocr if you want to run OCR yourself, pay only for compute, and get Markdown with layout from multilingual PDFs. Pick Google Document AI if you want a hosted API with per-page pricing, batch jobs and RAG chunking, and you already use Google Cloud. The main split is how you run it: a metered cloud service versus an MIT-licensed 3B model you host on your own GPU.
Google Document AIgoogle/document-ai | dots.ocrrednote/dots-ocr | |
|---|---|---|
| Type | OCR API | Open source |
| Price | 10 credits / page | 2 credits / page |
| Per 1,000 pages | $10.00 | $2.00 |
| Typical speed | 1.4 s / page | 1.5 s / page |
| Best for | Multilingual scans | Multilingual document parsing |
| Inputs | PDF, images | PDF, images |
| Languages | 200+ languages | 100+ languages |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | No | ✓ Yes |
| 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
- Enterprise Document OCR costs $1.50 per 1,000 pages and drops to $0.60 above 5M pages a month, with the first 1,000 free.
- Layout Parser includes chunking for RAG at $10 per 1,000 pages, so you do not build that step yourself.
- Batch requests accept up to 500 pages, which suits bulk OCR runs on Google Cloud.
Choose dots.ocr if
- There is no per-page fee. It is MIT-licensed code and you pay only for the compute you run it on.
- One model returns layout boxes, categories and text in one pass, with tables as HTML and formulas as LaTeX, plus Markdown output.
- The authors report 83.9 on olmOCR-bench for dots.mocr (v1.5), up from 79.1 for the original model.
Key differences
| Google Document AI | dots.ocr | |
|---|---|---|
| How you run it | Hosted API on Google Cloud | Self-hosted with vLLM or Hugging Face Transformers; a single CUDA GPU with vLLM is the documented path |
| Licence | Proprietary | MIT (the repository also includes a separate dots.ocr license agreement file) |
| Price | $1.50 per 1,000 pages for Enterprise Document OCR; $10 for Layout Parser; $30 for Form Parser | Free; you pay for compute |
| Output formats | JSON, plain text, bounding boxes, chunks | JSON, Markdown, HTML, LaTeX, bounding boxes, SVG |
| Request limits | Online requests capped at 15 pages (30 in imageless mode); batch up to 500 pages; Form Parser batch up to 100 pages | Not published |
| Table quality | 64.6% on Reducto's RD-TableBench, independent result, behind Azure and Textract | Authors say complex tables remain difficult at this model size; a table above 1000 grid cells raises an error and drops the whole page in Docling (issue #5146) |
| Known failure | Long PDFs do not fit in one synchronous call | Runs of ellipses or underscores can make output repeat endlessly; occasional parsing failures per the README |
Sources
- Document AI pricing
- Document AI limits
- Reducto RD-TableBench
- dots.ocr GitHub README
- dots.ocr blog / limitations
- Docling issue #5146
What each one is
Google Document AI
Google Cloud OCR with layout parsing and specialised processors.
Google Document AI detailsdots.ocr
Single vision model from rednote that does layout detection and content recognition together.
dots.ocr 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": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Google Document AI alternativesdots.ocr alternativesBest OCR models in 2026