Convert from PDF

Read a scanned PDF

Until now every converter here met a scan and said the same true, unhelpful thing: there are no characters in this file. This is the tool that does something about it — each page drawn, then read, on this device.

On this device

Choose a scanned PDF

The recogniser is downloaded the first time you use it, and a long document takes a few seconds a page.

    Opened on this device · never transmitted

    What this tool does, exactly

    What it does
    Draws each page as an image at 200 dpi, then recognises the text in it with Tesseract's LSTM engine, and joins the pages into one plain-text document. Both halves run in this browser tab; the scan never leaves the device.
    When it refuses
    When the PDF already has a text layer. Recognising a page whose characters are already there replaces perfect text with a guess — slower, worse, and invisibly so. The tool checks first, says how many characters it found, and sends you to PDF to Markdown or PDF to Word, which read what is really in the file.
    Why 200 dpi
    Because the recogniser wants about thirty pixels of x-height and falls off sharply below it, and ten-point body text at 150 dpi is under that line. 300 dpi buys very little more on printed text and costs a phone its memory, so it is the ceiling rather than the default.
    What it gives back
    The words in reading order, as plain text, with the confidence it had in each one. Not the layout: columns, tables, headers and footers are not reconstructed, and neither is the original as a searchable PDF — what you get is the text to paste somewhere else.
    What it costs to load
    About five and a half megabytes the first time — the page renderer, the recogniser and its English model — fetched only when you open this tool and cached by your browser afterwards. It is the heaviest thing on this site, which is exactly why nothing else pays for it.

    About this tool

    How is this different from PDF to Markdown or PDF to Word?

    Those two read the characters a PDF already contains, which is fast and exact. This one is for the documents that contain none — a scan is a picture of text — and it recognises the shapes instead. If your PDF has a text layer this tool will tell you so and send you to them, because reading what is there always beats guessing at it.

    How accurate is it?

    On a clean, flat, well-lit scan of printed type, high enough to paste and skim. On a phone photograph of a curved page in poor light, low enough that you should read it against the original. The tool reports its confidence for every word and lists the doubtful ones, so the checking has somewhere to start rather than being a guess about a guess.

    Do I get a searchable PDF back?

    No. You get the text. Producing a PDF with an invisible text layer laid over the original scan is a different and much larger job — it means writing the recognised characters back into the document at the exact position they were found. It is worth building and it is not built, and saying so is better than shipping something that looks like it and is not.

    How long does a long document take?

    Seconds per page, on one thread, plus the five-and-a-half-megabyte download the first time. A forty-page scan is minutes rather than seconds. Pages are drawn and read one at a time and released before the next, so memory stays flat — which is what makes a long document possible on a phone at all, if not quick.

    Which languages does it read?

    English only, in this version. Each additional language is about two megabytes on top of what is already the heaviest engine here, so they arrive as a choice you make rather than a cost everybody pays.

    Does the scan leave my device?

    No. The renderer, the recogniser, the language model and the recognition all live in this tab. This matters more here than almost anywhere on the site: the documents people scan are contracts, medical letters, payslips and identity papers, and every free online OCR service receives all of them.

    More convert from pdf tools · All 25 · How they work