Get the text out of any document
Photograph a bill, a form or a printed page — in Tamil, Hindi, Bengali or any of the eleven languages Kuraly supports — and get the text back as something you can edit, reshape and keep.
Bills and receipts
Photograph a shop bill and get the items, quantities and amounts back as a table you can work with.
Forms and applications
A filled-in form becomes a list of fields and values — names, dates, addresses, phone numbers.
Printed pages and letters
A page from a book, a notice, a printed letter. The text comes back as text you can edit and correct.
Four steps
Upload
A JPEG, PNG, WEBP or PDF, up to 10 MB. A PDF is read page by page.
Kuraly reads it
In the script it was written in. It runs in the background, so you can close the page.
Ask for the shape you want
"Make a clean invoice table", "list the fields and values" — in your own words.
Keep it
Editable, and saved into your workspace folders with everything else you write.
What we measured
Most tools claim to support a list of languages without saying what that means. We built 56 test documents across all eleven — printed pages, shop bills, filled-in forms, and deliberately degraded phone photographs of them — and measured what came back against the exact text we put in.
99.6%
of characters read correctly, across 56 documents in 11 languages
100%
of invoice references, dates, amounts and phone numbers came through intact
Under 10s
for a typical scan, including a three-page PDF
And what we have not measured
- Handwriting. Supported experimentally, but we have not measured it, so we are not going to tell you how well it works. Printed text is what the scanner is built for.
- Badly damaged documents. Our photo tests are skewed, soft-focused and dimly lit, but they start from a clean page. A crumpled receipt photographed in a dark room is a harder problem.
These figures are from our own testing, not an independent benchmark. How well it reads your document depends on how clearly it was printed and photographed.
Eleven languages, all of them tested
Every language below was measured on four kinds of document, not simply accepted as a setting.
Questions
Which file types can I upload?
JPEG, PNG, WEBP and PDF, up to 10 MB per file. A multi-page PDF is read page by page, and each page counts towards your allowance. A single image counts as one page.
How accurate is it?
In our own testing across 56 documents in 11 languages — printed pages, shop bills, filled-in forms, and phone photographs of them — 99.6% of characters were read correctly, and every invoice reference, date, amount and phone number came through intact. Accuracy on your own documents depends on how clearly they are printed and photographed.
Does it work with handwriting?
Handwriting is supported experimentally. You can tell Kuraly a document is handwritten and it will be processed differently, but we have not measured how well it performs on handwriting, so expect to correct more of the result. Printed text is what the scanner is built for.
How long does a scan take?
Usually under ten seconds. In our testing the median was under seven seconds, and a three-page PDF was no slower than a single photo. Scanning runs in the background, so you can close the page and come back.
Do I need an account?
Yes. The result is saved to your workspace as a document you can edit, so scanning needs an account. Writing and correcting text can be tried without one.
What does it cost?
Scanning is counted in pages. Every plan includes a monthly page allowance, and top-up packs cover anything beyond it. Reformatting text that has already been read does not cost another page.
Try it on something you have
A bill in your pocket, a form on your desk. Upload it and see what comes back.
Scan a document