Polish documents cross the desk of anyone handling UK settled-status cases, Polish citizenship confirmations, or genealogy-based passport claims: odpis (certified copies) of birth and marriage acts, criminal record certificates from the KRK, and old parish-registry extracts. Reglyph reads the scan with OCR + AI, removes the Polish from the page image, and sets the English translation into the same official form — so the USC office headers, act numbers, and stamps stay put.
Polish adds ą, ć, ę, ł, ń, ó, ś, ź, ż to the Latin alphabet, and OCR that misreads ł as l or ż as z corrupts surnames — a serious issue when the whole point of the document is proving identity or descent. Polish official documents also use rigid formal formulas ('odpis zupełny aktu urodzenia') whose distinctions matter legally: a complete copy is not an abridged one. Reglyph reads the full alphabet, keeps names spelled as issued, and preserves the act-number and registry layout that officials check first.


Drag to compare — every table, figure, and number stays in place.
These are the Polish-language documents people most often translate to English — each keeps its original layout, seals, and tables.
Drop in a scanned Polish PDF, a digital PDF, or a photo of the page. Image-only files are fine — no text layer needed.
OCR extracts the Polish text, then it's translated to English while the original text is erased from the page.
Get a clean PDF where tables, figures, stamps, and numbers sit exactly where they were.
Yes — upload the scan and the English version keeps the registry layout, act number, and annotations block exactly where the Polish original has them, with the distinction between a complete and abridged copy preserved in the title. UK and Irish authorities usually want such translations certified, so have a qualified translator sign off on the machine draft. The formatting work — normally the tedious half of the job — is already done.
On reasonably clear scans, yes — the OCR is built to read the full Polish alphabet rather than flattening Wałęsa to Walesa. For genealogy documents, be aware that many pre-war and parish records are handwritten, and handwriting is where OCR reliability genuinely drops; typed or printed registry copies work far better. When in doubt, order a typed odpis from the USC and translate that.
Upload the zaświadczenie and you'll have an English version in minutes, laid out like the original with the 'NIE FIGURUJE' (no record) determination in its place on the form. Five pages a day are free with no credit card, which covers a KRK certificate comfortably. If the employer or embassy requires a certified translation, the machine version serves as the draft the certifier reviews.
Agencies translate from the scan by hand over days; Reglyph rebuilds the English into the original layout in minutes, stamps and form boxes in place. For official submissions that require a certified or sworn translation, the draft is what the certifying translator reviews.
Simple, scan-friendly pricing. Pages are pages — no multiplier for scanned files.
$0.30 a page, from 10 pages — you pay for exactly 10 pages.
Vision models genuinely read scans now. The question isn't whether ChatGPT can translate your PDF — it's whether a chat reply is actually what you need back.
Read guideCamera-translate apps are great for reading a page on the spot, but they don't give you a file — here's how to get an actual translated PDF from your phone.
Read guideRandom symbols, blank pages, or 'PDF tools are temporarily offline' — the cause is almost always the same one thing, and it's fixable.
Read guide