Help centre

The practical details.

Everything you should know before trusting OCR.so with a document.

What can I upload?

PDF, PNG, JPEG, and WebP files up to the limit shown in your upload screen. A single document can contain up to 1,000 pages.

Do password-protected PDFs work?

Yes. PDF passwords are encrypted at rest, never returned or written to logs, and removed when the document expires or is deleted.

When are credits charged?

We reserve the page count or five credits, whichever is higher. A completed document costs at least five credits. On a partial result, only successful pages are charged; the minimum applies once the document eventually completes. Cancellation before processing costs nothing.

How long do you keep my files?

At most seven days. You can delete a document earlier at any time, which immediately removes its source, OCR artifacts, and result.

Do you train on customer documents?

No. Customer documents and extracted content are never used to train models.

What does the download contain?

One UTF-8 Markdown file in the original page order, with clear separators between pages. JSON and raw model diagnostics are not exposed.

Still need help?

Email [email protected]. Include the job ID, but never send a PDF password by email.