OCRRouter

Kraken vs Hunyuan OCR

Pick Kraken for historical manuscripts, early prints and right-to-left scripts such as Arabic and Hebrew, where you can train layout and recognition on your own material and need ALTO or PageXML output. Pick Hunyuan OCR for a single ~1B model that parses documents, spots text in photos and translates text in images, and can run on a laptop through llama.cpp. The main split is trainable archival tooling versus a ready-to-run vision-language model.

Krakenkraken/krakenHunyuan OCRtencent/hunyuan-ocr
TypeOpen sourceOpen source
Price1 credit / page2 credits / page
Per 1,000 pages$1.00$2.00
Typical speed1.5 s / page1.5 s / page
Best forHistorical documents and manuscriptsChinese documents and photos
InputsimagesPDF, images
LanguagesMany scripts, incl. right-to-left100+ languages
Max pages per request500500
TablesNo✓ Yes
Handwriting✓ YesNo
MathNo✓ Yes
Bounding boxes✓ Yes✓ Yes
ConfidenceNoNo
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 Kraken if

  • Your material is historical manuscripts or early prints, or uses 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.
  • You want an Apache-2.0 licence with no regional exclusions, and a public Zenodo repository of community-trained models to start from.

Choose Hunyuan OCR if

  • You need photo and scene text spotting, information extraction or text-image translation from one ~1B model.
  • You want Markdown, HTML, LaTeX or JSON output, with images up to 4K and a 128K context window in v1.5.
  • You work on Windows or a laptop, since the llama.cpp GGUF path runs on CPU or a consumer GPU, and you are outside the EU, UK and South Korea.

Key differences

KrakenHunyuan OCR
LicenceApache-2.0Tencent Hunyuan Community License. It excludes the EU, UK and South Korea, and a separate licence is needed above 100M monthly active users.
How you run itPython library installed with pip on Linux or macOS, x64 or ARM, Python 3.10 to 3.13. No Windows support.Self-hosted through vLLM, Transformers or llama.cpp. The unified vLLM environment recommends an NVIDIA GPU with CUDA 13. The llama.cpp GGUF path runs on CPU, consumer GPU or laptop.
Output formatsALTO, PageXML, hOCR, plain text, bounding boxesMarkdown, HTML, LaTeX, JSON, plain text, bounding boxes
TrainingLayout analysis, reading order and character recognition are trainable. Good accuracy usually needs training on your own material.One model covers parsing, spotting, extraction and translation, with 12 task types exposed via the vLLM client.
Scripts and languagesBuilt for historical and non-Latin scripts, including right-to-left, BiDi and top-to-bottom.Covers low-resource and ancient-script OCR among its task types.
BenchmarkNot publishedParseBench-2 overall 67.11 and ParseBench mean 63.86 (v1.5), with a text score of 62.88 in public CI replay.
Known issuesInstall can fail with a dependency conflict (issue #816). Learned line order is lost in segmentation and text output (issue #820).Needs CUDA 13 and vLLM 0.25.1+ for the unified environment. An archived 1.0 note said Transformers and vLLM outputs differed in accuracy.

Sources

What each one is

Kraken

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

Kraken details

Hunyuan OCR

Small end-to-end OCR model from Tencent covering parsing, extraction and photo translation.

Hunyuan 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": "kraken/kraken","fallbacks": ["tencent/hunyuan-ocr"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Kraken alternativesHunyuan OCR alternativesBest OCR models in 2026