Extract text from a PDF ยท copy ยท save as .txt
Pull the text out of a PDF to copy it all at once or save it as a .txt file. Documents that scramble when you drag-select in a viewer come out clean, page by page. Pick a range and merge lines into paragraphs.
๐ Your files are never uploadedCount the characters of the text, compare two versions with compare text, or remove duplicate lines with clean up text. To save pages as pictures use PDF to image.
Why copying from a viewer scrambles the text
A PDF stores text as "draw these characters at these coordinates". There is no notion of a line or a paragraph in the file, which is why drag-selecting in a viewer interleaves two columns or drags the page header into the middle of a sentence. This page reads the coordinates, gathers characters at the same height into one line, stacks the lines top to bottom and joins the pages in order. "Merge into paragraphs" then removes the forced break at the end of each line and keeps a blank line only where the gap between lines is wide โ between paragraphs.
If nothing comes out, it is a scan
A PDF made by scanning paper or photographing a page looks like text but is a picture. The file holds no characters, so no tool can extract them as they are; that takes OCR (optical character recognition), which guesses letters from their shapes and makes mistakes. This page counts such pages as "pages without text" and tells you when the whole file is like that. If you can get the original document (Word, Google Docs), it will always be more accurate than OCR.
Why it runs in the browser โ the PDFs people need text from are contracts, medical reports, bank statements and papers under review. Sending them to a conversion server is a risk you do not need to take: the extraction here happens on your device, works offline, and leaves nothing behind when the tab closes.
Frequently asked questions
Is my file uploaded?
No. The text is read out by your own browser; the page works with the internet switched off. Nothing is stored.
Tables come out broken
Table cells have positions but no cell boundaries in the text data, so the cells of one row are joined with spaces. Keep "Keep lines as shown" so at least the rows are right. To move a table into a spreadsheet, paste a row and use "Text to columns".
The characters are in the wrong order
Two-column layouts and vertical text put characters from different columns at the same height, so they can be merged into one line. Some PDFs also deliberately scramble their fonts to prevent copying; those cannot be extracted correctly by any tool.
Letters show as boxes or question marks
The PDF lacks the font mapping needed to turn glyphs back into characters. A PDF exported again with "embed fonts" from the program that made it will work. If someone sent you the file, only they can fix that.
How many pages can it handle?
Hundreds. Pages are read one at a time with a progress message. To extract only part of a document, enter a page range such as 1-3, 7.
Which encoding is the .txt?
UTF-8, which every modern editor and operating system reads correctly. If an old program shows garbled characters, open the file in Notepad and re-save it with a different encoding.