Amazon Textract vs Kraken
Pick Amazon Textract for English forms, invoices and large scanned PDF batches on AWS, where you want a managed API and per-page pricing. Pick Kraken for historical manuscripts, early prints and non-Latin or right-to-left scripts, where you can train on your own material and run it yourself on Linux or macOS. The main split is a hosted, proprietary service limited to six Latin-script languages versus a trainable, Apache-2.0 library built for scripts Textract does not cover.
Amazon Textractaws/textract | Krakenkraken/kraken | |
|---|---|---|
| Type | OCR API | Open source |
| Price | 25 credits / page | 1 credit / page |
| Per 1,000 pages | $25.00 | $1.00 |
| Typical speed | 1.2 s / page | 1.5 s / page |
| Best for | Forms and key-value extraction | Historical documents and manuscripts |
| Inputs | PDF, images | images |
| Languages | 6 languages | Many scripts, incl. right-to-left |
| Max pages per request | 3,000 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | ✓ Yes | ✓ Yes |
| Math | No | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | ✓ Yes | No |
| Forms | ✓ Yes | No |
| Multilingual | No | ✓ 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 Amazon Textract if
- You need tables, forms or queries from English documents, priced at $0.015, $0.05 and $0.015 per page for the first 1M pages a month.
- You process big scanned PDFs, since the async API accepts files up to 500 MB and 3,000 pages.
- You want word and line geometry with confidence scores returned as JSON from a managed API, with no model to train or host.
Choose Kraken if
- You work with historical manuscripts, early prints or Arabic, Hebrew and other right-to-left scripts, which Kraken supports along with BiDi and top-to-bottom text.
- You need to train layout analysis, reading order and character recognition on your own material, and can reuse community models from the public Zenodo repository.
- You need ALTO, PageXML, abbyyXML or hOCR output for digital humanities tools, under an Apache-2.0 licence with no per-page fee.
Key differences
| Amazon Textract | Kraken | |
|---|---|---|
| Licence | Proprietary | Apache-2.0 |
| How you run it | Hosted AWS API | Python library you install with pip on Linux or macOS (x64 or ARM, Python 3.10 to 3.13). No Windows support. |
| Price | Text detection $0.0015 per page; Tables $0.015; Forms $0.05; Queries $0.015 (US West Oregon, first 1M pages a month) | Free and self-hosted; you pay for compute |
| Languages and scripts | Text detection: English, French, German, Italian, Portuguese, Spanish. Handwriting and Queries: English only. Vertical Japanese and Chinese not supported. | Built for historical and non-Latin scripts, including right-to-left, BiDi and top-to-bottom |
| Output formats | JSON, plain text, bounding boxes | ALTO, PageXML, hOCR, plain text, bounding boxes |
| Table benchmark | 80.9% on RD-TableBench (independent) | Not published |
| Known failure | Sync API takes one page; no Markdown output without post-processing | Install can fail with a dependency conflict (issue #816); learned line order is lost in segmentation and text output (issue #820); usually needs training for good accuracy |
Sources
What each one is
Amazon Textract
AWS OCR with form key-value pairs, tables and per-word confidence scores.
Amazon Textract detailsKraken
OCR engine for historical documents and manuscripts, with trainable models for unusual scripts.
Kraken 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": "aws/textract","fallbacks": ["kraken/kraken"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Amazon Textract alternativesKraken alternativesBest OCR models in 2026