OCRRouter

RapidOCR vs Kraken

Pick Kraken for historical manuscripts, early prints and right-to-left scripts like Arabic and Hebrew, where you can train layout and recognition on your own material and export ALTO or PageXML. Pick RapidOCR for offline OCR of modern Chinese and English images on CPU, or when you need to embed OCR in a Java, C# or C++ app. The main split is target material: RapidOCR is a ready-to-run runtime for modern Chinese and English text, while Kraken is a trainable system for historical and non-Latin sources.

RapidOCRrapidai/rapidocrKrakenkraken/kraken
TypeOpen sourceOpen source
Price1 credit / page1 credit / page
Per 1,000 pages$1.00$1.00
Typical speed0.6 s / page1.5 s / page
Best forLightweight OCR on CPU or edgeHistorical documents and manuscripts
Inputsimagesimages
LanguagesChinese, English and moreMany scripts, incl. right-to-left
Max pages per request100500
TablesNoNo
HandwritingNo✓ Yes
MathNoNo
Bounding boxes✓ Yes✓ Yes
Confidence✓ YesNo
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

  • Your input is Chinese or English screenshots and receipts, which its default models cover.
  • You need to embed OCR in Python, C++, Java or C# and run it offline on CPU through ONNX Runtime.
  • You work on Windows, since Kraken installs on Linux and macOS only.

Choose Kraken if

  • You transcribe historical manuscripts or early prints, which is what it is built for.
  • You handle Arabic, Hebrew or other right-to-left, BiDi or top-to-bottom scripts.
  • You need ALTO, PageXML, abbyyXML or hOCR output for digital humanities tools, and want to train layout analysis, reading order and recognition yourself.

Key differences

RapidOCRKraken
Target materialChinese and English screenshots and receiptsHistorical manuscripts, early prints and non-Latin scripts
TrainingRuns converted PaddleOCR models; no training workflow describedTrainable layout, reading order and recognition; usually needs training on your own material for good accuracy
LanguagesChinese and English by default; other languages need extra modelsRight-to-left, BiDi and top-to-bottom scripts supported
Output formatsPlain text, bounding boxesALTO, PageXML, abbyyXML, hOCR, plain text, bounding boxes
Platform and bindingsPython, C++, Java and C#; CPU via ONNX Runtime by default, with optional GPU, TensorRT, OpenVINO and MNN enginesPython on Linux or macOS (x64 or ARM), Python 3.10 to 3.13; no Windows support
Licence and priceApache-2.0 (converted models keep their upstream licences); free, self-hostedApache-2.0; free, self-hosted
Known failure16-bit grayscale PNG or TIFF give no text or wrong text (issue #776); Arabic recognition crashes because python-bidi is undeclared (issue #719)Install can fail with a dependency conflict (issue #816); learned line order is lost in segmentation and text output (issue #820)

Sources

What each one is

RapidOCR

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

RapidOCR details

Kraken

OCR engine for historical documents and manuscripts, with trainable models for unusual scripts.

Kraken 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": ["kraken/kraken"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

RapidOCR alternativesKraken alternativesBest OCR models in 2026