Shot2Text

Scanned PDF to Text — Read Image-Only PDFs Free Online

Scanned PDF to text, because a scan is not a document — it is a set of photographs of one, which is why selecting text in it returns nothing. This renders each page as an image and reads the words out, up to 20 pages at a time.

Only the language you pick is downloaded.

What is scanned PDF to text?

There are two kinds of PDF and they behave completely differently under the cursor. A digital PDF stores the letters themselves plus instructions for drawing them, so selecting text works exactly as it does in a word processor. A scanned PDF stores a photograph of each page. There are no letters in the file at all — just pixels. Selecting text in one selects nothing, because there is nothing there to select.

You can tell which you have in about two seconds: open the file and try to select a word. If the text highlights, it is a digital PDF and your PDF reader's copy button is the right tool — it gives you perfect, exact text. If nothing highlights, it is a scan, and the letters have to be recovered from the picture.

That recovery is what this page does. It renders each page to an image with the PDF engine, then runs the same text reader used everywhere else on this site over that image. The result is never perfect — recovering letters from pixels is guesswork where copying letters from a file is not — but it turns an unselectable document into something you can search, quote and paste.

What can you use scanned PDF to text for?

  • Scanned contracts and agreements. The version someone sent you is usually a scan. Get the clauses into text so you can search them and quote accurately.
  • Old statements and invoices. Banks and suppliers send PDFs that are pictures. Line items and totals come out in reading order.
  • Photocopied pages passed around an office. A photocopy that was then scanned is a photo of a photo — lower quality, but usually still readable if the original was clean.
  • Academic papers from a repository. Older papers are often image-only scans. Good enough to search; check anything you plan to quote.
  • Government forms and letters. Usually clean printed text on plain paper, which reads well. Reference numbers need checking.
  • Books and manuals. Multi-column layouts come out interleaved because the reader scans line by line across the page — crop each column if that happens.

Why do scanned PDFs defeat a reader?

Scans are the worst-case input for text recognition, because they stack several kinds of damage on top of each other.

  • The scan is skewed Pages fed into a scanner at a slight angle produce diagonally running lines, and the reader loses whole rows toward the edge. Use your scanner's deskew option if it has one, or scan again with the page squared against the edge.
  • Faint photocopy of a photocopy Each generation loses contrast and adds grain. After two or three the difference between ink and paper is gone. Find a better original if the content matters. There is no recovery from a scan that has no contrast left.
  • Multi-column layout The reader follows lines across the full page width, merging the end of the left column with the start of the right. Crop one column per page and convert separately, or accept interleaved text and reorder by hand.
  • Password-protected PDF Encrypted files cannot be opened for rendering without the password. Remove the protection in your PDF reader first, then upload the unprotected copy.
  • Tables and forms The reader returns words but has no concept of cells, so a table becomes a stream of values with the column structure lost. Expect to rebuild the table. The values are there, in roughly left-to-right order per row.
  • Handwriting on a scan Same limits as handwriting anywhere — printing converts, cursive does not. Use the handwriting tool for pages that are mostly handwritten.
  • Watermarks and stamps Overlaid marks are read as characters and injected into the middle of words. Usually unavoidable. Expect stray characters where the watermark crosses text.

How many pages can it handle, and how fast is it?

Up to 20 pages per PDF. Longer files are truncated with a message rather than failing silently — split the document and convert the rest in another pass.

Each page is rendered at roughly twice the PDF's nominal resolution, which is about 144 dpi. That is a deliberate compromise: high enough for the reader to resolve normal body text, low enough that a 20-page document does not exhaust the memory of a phone.

Pages are separated by a short horizontal rule in the output so you can find where one page ends. Text appears in reading order per page; there is no attempt to preserve layout, columns or tables.

Speed depends on page count: roughly two to four seconds per page after the initial 3 MB of English data has downloaded. A 20-page scan takes a minute or two. Keep the tab in the foreground — browsers throttle background tabs, which pauses the work.

Why are scanned documents the sensitive case?

Contracts, bank statements, medical letters, employment paperwork — the documents that arrive as scans are usually the ones you would least want on a stranger's server. Every free online converter asks you to upload exactly these.

Here nothing is uploaded. The PDF is parsed and each page rendered inside your browser tab, and the reading happens there too. The file never travels. There is no upload button because there is no upload. You can confirm it: open your browser's developer tools, watch the network panel during a conversion, and you will see the PDF engine and language data download — and no request carrying your document.

Questions

Why does copy and paste give me nothing from my PDF?
Because it is a scanned PDF. A scan stores a photograph of each page rather than the letters, so there is no text in the file to select — only pixels. You can confirm it by trying to highlight a word: if nothing highlights, it is a scan and the letters have to be recovered from the image, which is what this tool does.
How do I know if my PDF is scanned or digital?
Open it and try to select a word. If it highlights, the PDF is digital and you should just copy from your reader — that gives you exact text with no errors. If nothing highlights, it is scanned. There is no advantage to running a digital PDF through OCR; extraction from the file is always better than guessing from pixels.
How many pages can it handle?
20 per PDF. Longer documents are truncated after page 20 with a message rather than failing — split the file and run the rest in another pass. The limit exists because each page is rendered to a full-resolution image, and rendering hundreds of pages would exhaust memory on most devices.
Why is it slow compared to the image tools?
Because there are two stages per page instead of one: the PDF engine has to render the page to an image, then the text reader has to read that image. Budget two to four seconds per page once the English data has loaded. Keep the tab in the foreground, since browsers throttle work in background tabs.
Does it preserve the layout, columns and tables?
No. It returns words in reading order per page. Multi-column documents come out interleaved because the reader scans line by line across the full page width, and tables become a stream of values with the structure lost. You get the content, not the formatting.
Is my PDF uploaded?
No. The file is read in your browser tab — parsed there, rendered there, recognised there. Nothing is transmitted and nothing is stored. You can verify this in your browser's network panel: no request carries your document.
It says my PDF is password protected. What now?
Encrypted files cannot be rendered without the password. Open the PDF in your reader, unlock it, and re-save or print-to-PDF an unprotected copy. Then upload that. Your reader will handle the decryption; this tool cannot.
Can it read handwriting in a scanned PDF?
Partly, with the same limits as handwriting anywhere: neat printing converts, joined-up cursive usually does not. If a scanned page is mostly handwritten, the dedicated handwriting tool gives better guidance and applies contrast correction tuned for photos of paper.
Are scanned PDFs in other languages supported?
18 languages are available from the dropdown above the tool. Numbers, URLs and email addresses survive in any of them, but pick the right language before converting or words will come out wrong.