General
Does Kaizen OCR work offline?
Yes — Simple OCR, Advanced OCR, all PDF tools, all image tools, screenshot capture, and hotkeys work entirely offline. The only feature that requires internet is Searchable PDF (which uses your own Azure Document Intelligence resource).
Is my data sent anywhere?
No, with one exception: if you use Searchable PDF, the document is sent directly to your Azure endpoint — not through our servers. Your Azure key is stored encrypted in a local SQLCipher database and never leaves your machine.
Which OCR engine should I use?
Start with Paddle AI — fast and accurate on most modern documents. Switch to Tesseract for older scans or specialty languages. Use Searchable PDF (Azure) for the best accuracy on complex layouts, tables, or handwriting.
Pricing & licensing
What does “7 ops per feature” mean?
Each feature has its own quota. You can run Simple OCR 7 times, Advanced OCR 7 times, Create PDF 7 times, etc. — all on Free. Screenshot capture and hotkeys are unlimited on Free.
Do unused Free-tier operations expire?
No. They stay forever.
What's the difference between Pro and Premium?
Both lift the Free-tier quotas and unlock Searchable PDF. Pro is a yearly license; Premium is a lifetime license (one-time purchase).
Can I use my license on multiple machines?
One machine at a time. Deactivate on one machine to free up the slot, then activate on another. No paperwork.
Is there a refund policy?
3-day no-questions-asked refund. Within 7 days, case-by-case. Details here.
Azure & Searchable PDF
Does Searchable PDF cost money?
Yes, but it's paid to Microsoft, not us. Azure's F0 tier gives you 500 pages/month free; S0 is pay-as-you-go at $1.50/1000 pages (Read) or $10/1000 (Layout).
What's the setup for Azure?
We have a 7-step guide with screenshots.
Can I use a different OCR cloud service?
Not currently — Azure Document Intelligence is the only cloud backend supported for Searchable PDF. If there's demand for AWS Textract or Google Document AI, let us know.
Files & formats
What file formats can Kaizen OCR read?
See the full format reference. In short: PNG, JPG, BMP, TIFF, GIF, WEBP, and PDF.
Can it handle huge PDFs?
Yes. PDFs are processed page-by-page, so a 500-page document just takes proportionally longer — it doesn't load into RAM all at once. For Azure Searchable PDF, very large documents may need splitting first.
Does it read handwriting?
Sometimes. Neat printing with Paddle AI works. Cursive or messy handwriting needs Azure Document Intelligence — it's far ahead of local engines for handwriting.
Troubleshooting
OCR is slow
Paddle AI uses AVX2 CPU instructions when available — machines without AVX2 (pre-2013) run 2×–5× slower. Check if you have AVX2 by looking at Settings → OCR engine. Also, Advanced OCR with heavy preprocessing (denoise + deskew + threshold) takes longer than no preprocessing.
Hotkeys aren't working
Make sure Enable global hotkey is ticked in Settings. Some hotkey combos conflict with Windows or other apps — try a different combo. See the Screenshot capture guide.
Confidence scores are low
Usually means the source image is low quality. Try: (1) scanning at higher DPI (300+), (2) running Image Deskew first, (3) switching to a different OCR engine, or (4) using Searchable PDF (Azure) for the hardest documents.
I can't see my problem here
See the troubleshooting guide, or email [email protected].