OCRRouter

RapidOCR vs Google Cloud Vision OCR

Pick RapidOCR if you want free, offline OCR inside your own app, mainly for Chinese and English screenshots and receipts. Pick Google Cloud Vision OCR if you need hosted, high-volume text extraction with handwriting support and a word-level JSON hierarchy. The main split is where it runs: RapidOCR on your own CPU under Apache-2.0, Vision as a paid API with no on-premise option.

RapidOCRrapidai/rapidocrGoogle Cloud Vision OCRgoogle/cloud-vision
TypeOpen sourceOCR API
Price1 credit / page2 credits / page
Per 1,000 pages$1.00$2.00
Typical speed0.6 s / page1.0 s / page
Best forLightweight OCR on CPU or edgePhotos and simple scans
InputsimagesPDF, images
LanguagesChinese, English and more100+ languages
Max pages per request1002,000
TablesNoNo
HandwritingNo✓ Yes
MathNoNo
Bounding boxes✓ Yes✓ Yes
Confidence✓ Yes✓ Yes
FormsNoNo
Multilingual✓ Yes✓ 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 RapidOCR if

  • It is free and self-hosted under Apache-2.0, so you pay only for compute and your images never leave your machine.
  • It runs offline on CPU through ONNX Runtime, with optional TensorRT, OpenVINO or MNN engines, and embeds in Python, C++, Java or C# apps.
  • It suits Chinese and English screenshots and receipts, and it is available as an OCR backend option in Docling.

Choose Google Cloud Vision OCR if

  • It supports handwriting through DOCUMENT_TEXT_DETECTION with language hints, which suits handwritten notes.
  • It returns a page, block, paragraph, word and symbol hierarchy with bounding polygons, so you get structure down to single characters.
  • It scales without servers: async batch takes up to 2,000 image files per request, and the price falls to $0.60 per 1,000 units above 5M.

Key differences

RapidOCRGoogle Cloud Vision OCR
How you run itLibrary you run locally in Python, C++, Java or C#Hosted API over REST or client libraries on Google Cloud
LicenceApache-2.0 (converted models keep their upstream licences)Proprietary
PriceFree, self-hosted; you pay for computeFirst 1,000 units per month free per feature, then $1.50 per 1,000 units up to 5M, $0.60 above; each PDF page is one unit
OutputPlain text and bounding boxesJSON, plain text and bounding boxes, with page, block, paragraph, word and symbol hierarchy
Languages and handwritingChinese and English only by default; other languages need extra modelsHandwriting supported via DOCUMENT_TEXT_DETECTION and language hints
Tables and layoutNot for layout, tables or Markdown outputNo table or layout structure and no Markdown output; Google points to Document AI for forms and tables
Known failure16-bit grayscale PNG or TIFF give no text or wrong text; Arabic recognition crashes because python-bidi is undeclaredBilling must be enabled even within the free 1,000 units; PDF and TIFF need async processing via Cloud Storage

Sources

What each one is

RapidOCR

PaddleOCR models packaged for ONNX Runtime and other engines. Easy to deploy without Paddle.

RapidOCR 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": "rapidai/rapidocr","fallbacks": ["google/cloud-vision"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

RapidOCR alternativesGoogle Cloud Vision OCR alternativesBest OCR models in 2026