Extract Text from a PDF
Get the words out of a PDF as plain text you can copy and edit. Note that scans have no text layer — there is nothing to extract from an image.
How to use it
- Drop the PDF into the dashed box.
- Press Extract text; a few seconds later the text appears in the box below.
- Copy it directly, or download it as a .txt file.
- For long documents, tick the page-marker option so you can tell which page a passage came from.
What is actually being extracted
There are two kinds of PDF. A text PDF stores every character with its coordinates plus a table mapping
glyph codes back to characters. An image PDF stores a picture and no characters at all. This tool reads
the first kind: it pulls characters by coordinate, starts a new line when the vertical position changes,
and inserts a space where the horizontal gap is wide.
So it returns content but not layout. Columns may be spliced together
horizontally, a table becomes rows of text that no longer line up, and headers or footers may end up in
the middle of a paragraph. That is the limit of plain text, not a failed extraction.
Why scans produce nothing
A scanner photographs paper and outputs pixels; not one character is machine-readable. Some scanning
software runs OCR and pastes the result on as an invisible text layer — those files extract fine. Files
that never went through OCR return nothing.
This site does not do OCR. A recognition model is large, slow on first load and inconsistent on
imperfect scans, and offering a weak version would waste your time. For scanned documents, a dedicated
OCR tool is the honest recommendation.
After extracting
The text stays in your browser. Nothing is uploaded, nothing is sent to a model, no server sees it. If
you need it translated, rewritten or summarised, copy it into whichever tool you use for that. The
downloaded .txt carries a UTF-8 BOM so that spreadsheet software and older Windows editors detect the
encoding correctly.
Why is the output empty?
Because the PDF has no text layer — it is a stack of images, which is what you get from a scanner, a phone photo or an image-only export. A text layer is invisible characters stored alongside the pixels; a scanner records only pixels, so there is nothing to extract. This site does not do OCR: a recognition model is tens of megabytes, slow to load and unreliable on poor scans, and a half-working version would mislead people.
Why is the extracted text out of order?
Because PDF has no concept of a paragraph — only text runs, each with coordinates. The tool infers reading order from those coordinates: group by vertical position into lines, then sort horizontally. Columns, tables, captions, headers and footers all break that inference. Two columns can be spliced together line by line, and a footer can land in the middle of a paragraph. Every plain-text extractor has this problem; it is not a damaged file.
Can it keep the original layout?
Plain text cannot carry layout, only line breaks. Preserving font sizes, indents and table rules means converting to Word or HTML, which requires layout analysis and a server-side model — not something a browser tab can do. That is exactly why this site does not offer PDF to Word.
Does it work with non-Latin text, accents and ligatures?
Yes, provided the file contains a proper ToUnicode mapping — the table that maps glyph codes back to characters. Normal exports include it. A few old files built with unusual font subsets extract as garbage or boxes, and no tool-side fix exists for that; you need the original document.
Why does the downloaded .txt start with an invisible character?
That is a UTF-8 BOM, a marker telling software that the file is UTF-8. Spreadsheet applications and some older Windows editors use it to detect the encoding correctly; modern editors such as VS Code handle the file either way.
How many pages can it process?
Several hundred is fine; speed depends on how much text there is. Extraction reads the text layer page by page, so with a long document let the progress bar finish rather than closing the tab.
Can it translate or summarise the text?
No. This tool extracts and stops there — everything stays in your browser and nothing is sent to any model or server. Copy the text into whichever tool you use for translation or summarising.
Can it extract from an encrypted PDF?
No. The text layer cannot be read until the file is decrypted, and this site does not remove passwords.