Live
OCR a PDF
Make scanned PDFs searchable
PDF to Text
LiveExtract text from a PDF into a plain .txt, with a marker above every page. Whole file, or only the pages you name.
What it does
- Whole document, or just the pages you name
- Page markers so you can tell where a line came from
- Says so plainly when a PDF is a scan, instead of returning an empty file
Your files
Gone within the hour.
Files are uploaded over HTTPS, used only to run this tool, and deleted from our server automatically about an hour later. Never sold, never used to train anything.
How to extract the text from a PDF
Add the PDF
One file, up to 50 MB. The pages render so you can confirm you have the right document.
Choose the pages, or leave it blank
Blank means the whole file. Otherwise the box takes ranges like 1-3, single pages like 5, and open-ended ones like 8-. Up to 500 pages in one job.
Press Extract text
This reads the text the PDF already stores, so it is fast and exact: what comes out is what is in the file.
If it comes back empty, it is a scan
A scanned page is a picture and stores no text at all. Rather than hand you an empty file, the tool says so and points you at OCR PDF, which adds the text layer first.
Frequently asked questions
Why did it say there is no text to extract?
Because the pages are pictures rather than text, which is what a scan or a photographed document is. The test is an average of 8 characters per page across the pages you asked for, so a stray character from a header stamp does not count as text: a 40 page scan that yields one character is refused rather than handed back nearly empty. Run OCR PDF on it first to add a text layer, then extract the text from that result.
Does the .txt keep the layout?
No. A plain text file has no columns, tables, fonts or images. If you need the layout, convert to Word instead.
Can I extract only some pages?
Yes. Type ranges like 1-3, 7, 12- in the pages field, where 12- means page 12 to the end. The pages come back in document order however you type them, and a page you name twice appears once: 10-12, 2-3 gives pages 2, 3, 10, 11 and 12, and 5, 5, 5 gives page 5 on its own. Leaving the box blank takes the whole document. The 500 page limit counts the pages you selected rather than the length of the file, so 100 pages out of a 900 page PDF is fine.
Will the reading order be right?
For a single column document, yes: the text is sorted into reading order rather than the order the file happens to store it in. Two columns are the case to know about. Where both columns sit on the same baselines, which is how most typeset two column pages are built, each line of the output carries the left column and then the right one joined by a run of spaces, so the columns read across instead of down. Where the columns are offset by half a line they alternate instead. No words are lost either way, but if you need one column and then the other, separate the columns before extracting.
What does the text file look like?
A UTF-8 .txt named text.txt, with a line reading --- Page 8 --- above each page's words. The number is the page's own number in the original document rather than its place in your selection, so asking for 8-10 gives markers for 8, 9 and 10. In a file that mixes typed pages with scanned ones the scanned pages still get their marker with nothing underneath it, which is how you tell which pages need OCR.
Can I extract text from a password protected PDF?
No. A PDF that needs a password to open is refused before any text is read, and the message points you at Unlock PDF: unlock a copy first, then extract from that. Owner restrictions that do not stop the file opening are a different thing and cause no trouble here. One PDF at a time, up to 50 MB.