Extract text from PDF
Runs on your device- Instant preview and copy
- Download as TXT
- Lines or flowing paragraphs
- Files never uploaded
or drop it here · up to 200 MB
Processed on your device. Nothing is uploaded.
Pull every word out of a PDF as plain text. Preview it, copy it with one tap, or save a UTF-8 TXT file, all without uploading anything. Scanned PDFs need OCR first, because this tool reads the existing text layer.
How it works
How to extract text from a PDF
- 1
Choose a PDF
Pick a PDF up to 200 MB. Text extraction starts as soon as the file opens, and a word and character count appears with the preview.
- 2
Pick a layout
Keep As laid out for line-by-line text, or Flow paragraphs to join lines into paragraphs for pasting into a document.
- 3
Copy or download
Use Copy all to paste the text anywhere, or Download TXT to save a file that opens correctly in Notepad, TextEdit and code editors.
Convert from PDF
About PDF to Text
Last updated
Why extract plain text instead of converting the PDF?
Plain text is the cleanest input for translation tools, AI assistants, word counters, search indexes, note apps and code. It strips out fonts, layout and images, leaving content that pastes anywhere. Extracting text is also the quickest way to check whether a PDF is searchable at all, or to recover wording from a document that can't be edited.
The text also shows exactly what a search engine, screen reader or document search tool will see. Before sending a PDF that has been redacted, extract its text here: if the blacked-out words still appear in the preview, the redaction only covered them visually, and Redact PDF should be used to remove them for real.
What is the difference between As laid out and Flow paragraphs?
As laid out keeps one line of text per line on the page, while Flow paragraphs joins each paragraph into a single line. The first suits addresses, invoices, tables, poems and anything where line breaks carry meaning. Wide gaps between words become tab characters, which helps columns line up again in a spreadsheet.
Flow paragraphs also removes hyphens that split words at line ends. It is the better choice for pasting into Word, Google Docs, an email or a translator, where hard line breaks would otherwise break every sentence apart. Switching between the two updates the preview instantly, without reading the PDF again.
| Mode | Line breaks | Best for |
|---|---|---|
| As laid out | One per line on the page; tabs for wide gaps | Invoices, tables, addresses, poems, code |
| Flow paragraphs | One per paragraph; end-of-line hyphens removed | Word, Google Docs, email, translators, AI tools |
How is the text rebuilt from the PDF?
The tool reads positioned text fragments with pdf.js and reassembles them into lines and paragraphs. PDFs don't store sentences; they store small pieces of text with coordinates. The pieces are sorted top to bottom and left to right, merged along the same baseline, and separated by spaces or tabs based on the gaps between them.
Paragraphs are separated using line spacing, indentation and sentence endings. Pages are processed one at a time, so documents with hundreds of pages work without freezing the browser. The same analysis feeds PDF to Word, which adds heading and list detection and writes a formatted DOCX instead of plain text.
Does it work on scanned PDFs?
No. A scanned page is an image, so it has no text layer to read, and this tool does not run OCR. The preview says when no text was found. OCR PDF recognises the words in the browser and adds an invisible text layer; the OCR result can then be opened here to extract or check the recognised text.
What are the limits of text extraction?
The output contains exactly what is in the text layer, in reading order across the page, with no formatting. That is usually what is wanted, but it has consequences for multi-column documents, hidden text and oddly encoded fonts that are worth knowing before relying on the result:
- Scanned pages have no text; use OCR PDF first.
- Text hidden behind images or outside the visible page can still appear, because it is part of the text layer.
- Headers, footers and page numbers are included as ordinary lines of text.
- Multi-column layouts are read across the page, so columns may interleave.
- Some PDFs made with unusual font encodings produce garbled characters; OCR PDF is the workaround.
- Need headings and bold kept? PDF to Word produces a formatted DOCX instead.
FAQ
PDF to Text questions
How do I copy text from a PDF that won't let me select it?
Open it here. The text layer is read directly, so it appears in the preview where Copy all works, even when a viewer makes selection awkward. An empty preview means the PDF is a scan and needs OCR PDF first.
Why is no text found?
The PDF contains pictures of pages, typically from a scanner or phone camera, not real characters. OCR PDF can recognise the words and add a text layer on the device, after which extraction works normally.
Is the extracted text stored anywhere?
No. Text is extracted by pdf.js inside the browser tab and exists only in that tab's memory until it is copied or downloaded. Nothing is sent to a server, and closing the tab discards the document and the text.
Will accented letters and other languages work?
Yes. The TXT file is saved as UTF-8 with a byte-order mark and Windows line endings, so accents, Cyrillic, Greek, Chinese and other scripts display correctly, including in Windows Notepad and older editors.
Why do table columns run together?
Plain text has no columns. Wide gaps between words on the page are turned into tab characters, so pasting the text into Excel or Google Sheets often lines the columns up again without any manual splitting.
Can I mark where each page starts?
Yes. Tick Mark page breaks to insert a separator line with the page number before each page's text. That is useful for citing page numbers or for splitting the text into per-page chunks later.
Keep going
Related tools
Got a PDF to fix?
Free, no account, no watermark. Most tools run on your device, so files never leave it.
Browse all tools