OCRRouter

MonkeyOCR vs Google Cloud Vision OCR

Pick MonkeyOCR for English and Simplified Chinese PDFs such as academic papers with formulas and tables, when you can self-host on a GPU and the licence fits. Pick Google Cloud Vision OCR for photos, handwriting and high-volume plain text where you want a hosted API and no GPU. The main split is structure versus plain text: MonkeyOCR returns Markdown and JSON with layout, while Cloud Vision returns text and boxes with no table or layout structure.

MonkeyOCRmonkeyocr/monkeyocrGoogle Cloud Vision OCRgoogle/cloud-vision
TypeOpen sourceOCR API
Price2 credits / page2 credits / page
Per 1,000 pages$2.00$2.00
Typical speed1.5 s / page1.0 s / page
Best forChinese and English PDFsPhotos and simple scans
InputsPDF, imagesPDF, images
LanguagesChinese, English100+ languages
Max pages per request5002,000
Tables✓ YesNo
HandwritingNo✓ Yes
Math✓ YesNo
Bounding boxes✓ Yes✓ Yes
ConfidenceNo✓ Yes
FormsNoNo
MultilingualNo✓ Yes
On OCRRouterAvailableAvailable

Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.

Which to choose

Choose MonkeyOCR if

  • You need Markdown, JSON, LaTeX or HTML from English or Simplified Chinese PDFs, including formulas and tables.
  • You want to run it yourself: it works on an RTX 3090/4090, A6000, A100 or H800, and on an 8GB RTX 4060 with quantized weights.
  • You are doing research or evaluation, where the authors report pro-3B at 0.138 / 0.206 overall edit distance and 81.5 / 87.5 table TEDS on OmniDocBench (EN/ZH).

Choose Google Cloud Vision OCR if

  • You need to read handwriting or photos of signs and labels, which MonkeyOCR does not fully support.
  • You want a hosted API with no GPU to run, with the first 1,000 units per month free and $0.60 per 1,000 pages above 5M units.
  • You need plain text extraction at volume, with a page, block, paragraph, word and symbol hierarchy and bounding polygons.

Key differences

MonkeyOCRGoogle Cloud Vision OCR
How you run itSelf-hosted via LMDeploy or Docker; Docker on 20/30/40-series and V100 needs a patched image to avoid a shared-memory errorHosted API over REST or client libraries on Google Cloud; billing must be enabled even for the free tier
Licence and commercial useCode Apache-2.0; v1 weights are research and non-commercial only, commercial use needs a written licenseProprietary service
PriceFree for research when self-hosted; commercial use needs a separate license from the authorsFirst 1,000 units per month free per feature, then $1.50 per 1,000 units up to 5M, $0.60 per 1,000 above; each PDF page is one unit
Output formatsMarkdown, JSON, LaTeX, HTML, bounding boxesJSON, plain text, bounding boxes; no Markdown, so you rebuild reading order yourself
Tables and layoutTable TEDS of 81.5 (EN) / 87.5 (ZH) for pro-3B, as reported by the authorsNo table or layout structure; Google points to Document AI for forms and tables
Handwriting, photos and languagesNot fully supported: photographed text, handwriting, Traditional Chinese and multilingual textHandwriting supported via DOCUMENT_TEXT_DETECTION with language hints; TEXT_DETECTION is for photos
Speed and batch limitsPro-3B about 0.49 pages/s on a 3090; pro-1.2B is about 36% faster for about 1.6% lower accuracyAsync batch accepts up to 2,000 image files per request; PDF and TIFF need async processing via Cloud Storage

Sources

What each one is

MonkeyOCR

Open-source document parser that splits work into structure, recognition and relation steps.

MonkeyOCR details

Google Cloud Vision OCR

Google Cloud's image OCR with text and document text detection, word boxes and confidence.

Google Cloud Vision OCR details

Try both on your own file

Run the file once with each model and compare the output.

PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?

Use one, fall back to the other

{"model": "monkeyocr/monkeyocr","fallbacks": ["google/cloud-vision"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

MonkeyOCR alternativesGoogle Cloud Vision OCR alternativesBest OCR models in 2026