PDF to text: convert a scanned or digital PDF
Select a word in your PDF. If it highlights, the text is already there. If not, the file is a picture of text, and someone has to read it. Or run pdftotext on it: if words come out, there is a text layer.
Pick one page of your PDF, scanned or digital, and read it free: text, with a box and a confidence for every word.
Word One word, Northlight. Box measured on the image.
Does a PDF carry a text layer? Check before you run OCR
A PDF with a text layer
The words are already in the file, and pdftotext shows them. The service reads the page image anyway, so the answer has the same shape as for a scan.
A scan, photo or fax as PDF
No text layer at all.
The service reads the page image of every page, so a digital PDF and a scan get the same kind of answer: text with a box and a confidence for every word.
PDF to text, Markdown or JSON with word boxes
Every word comes with text, box and confidence; Markdown and text give the words in reading order. A page that could not be read is named in the header X-Pages-Failed, a page that was not run in X-Pages-Not-Run; an empty page in a text result is then an unread page, not a blank one.
| You want | Format |
|---|---|
| Only the words | text |
| Headings and tables as Markdown | md |
| Words with box, confidence and origin | json |
How accurately OCR reads the numbers of a PDF
On four hard pages with a text layer in the PDF, the text layer served as the truth for the numbers.
Source: own run on four hard pages, reading model alone.
Which PDFs can I send?
- A password-protected PDF is answered with
password_required; send it without the password and it is read. - A document of up to 100 MiB and 1,000 pages goes through the API, as PDF, PNG, JPEG or multi-page TIFF. Word, Excel and HTML files are exported to PDF first.
- Broken input is refused with status 400, 415 or 422 and costs nothing. A TIFF and a page of 40 megapixels are read. After a hard kill of the service during a job, all 30 of 30 pages were charged exactly once. Source: robustness run on one service.
Contracts and statements stay in your own network when PDF to text runs on your own server.
Try PDF to text free on your own page
Choose one page of your own PDF and read it free, or open the OCR demo gallery to see scanned and digital PDFs read word by word.
Production use, high volume or your own server?
Private use and tests are free. For a company in production, a server of your own or a pipeline built on your pages, send the kind of page and the volume, and get a written offer, or a clear no when the fit is not there.
