3 free · then $7 for every tool

A scan has no text. Read the words anyway.

Each page is drawn in this tab and handed to a recognition engine running in the same tab. The result is the machine-read text, page by page, downloadable as a plain text file.

The reading
Free allowance

Pages that arrived as photographs of paper are drawn in this tab and read by an engine running in the same tab. Nothing is uploaded.

Remove the limit — $7
What this tool does not do
  • The download is text, not a searchable PDF — the pages of your document are not changed
  • Up to twenty pages are read in one run; any beyond that are reported as not read rather than silently dropped
  • Recognition misreads names, digits and dates. Treat the output as a draft to check against the page, not as a transcript
  • The language data shipped with this site is English only, so other scripts come out as noise
  • The first run downloads about 11 MB of engine, which is why it is slower than the runs after it