Scanned documents and OCR

1. The scanned source

   The file as you uploaded it — a picture of pages.

2. The reconstructed document

   The text is recognised, the reading order is reconstructed, tables become tables again.

3. The translated document

   From here it is an ordinary document, and it returns the same nine files.

The recognition stage runs on our own servers. The scanned page is not sent to any external provider in order to be read.

Scan quality — shown before payment

After the analysis and before you pay, the system shows you which of four categories your document falls into.

• Digital
• Good scan
• Complex scan / photo
• Low quality

At low quality: text recognition is less reliable, and some pages may need review by you.

Beside the rating you are shown the page count, an estimate of the word count, and how many pages are digital and how many are scanned. That is the proof the document was actually analysed — not a figure produced from nothing.

One order, one price

A scanned document is priced as two lines in the same quote: a page line for turning the scan into a document, and a word line for the translation itself. Both are shown before payment, and there is no second invoice afterwards.

The page line: 39¢ per page to start, declining to 29¢, 22¢ and 16¢ as the document gets longer — marginal bands.

A full worked example on the pricing page

What a scan does not become

• Handwriting is not recognised.
• A low-quality scan is not reconstructed perfectly — the rating exists precisely so you know that in advance.
• The system aims to preserve the source's structure, tables and reading order.
• Up to 200 pages per document · Up to 50MB per file

You don't have to translate. If all you need is for the scan to become an editable Word file or an Excel table — that is an order in its own right, with no translation line. Conversion services

Upload the scan, see what came out of it

The rating, the page count and the price all arrive before payment.