How to use Extract Text from PDF
- Choose one PDF with selectable text.
- Leave the page field blank for the full document or enter pages such as 1-3, 7.
- Extract, preview, copy or download the text file.
What this tool changes
PDF.js reads the existing text layer from each selected page. PDF North keeps the page sequence, creates a plain-text preview and TXT download, and reports page, word and character counts.
From the makers of PDF North
Lost your tabs after a Chrome crash or update? TabRollback backs up your open tabs so you can restore them in one click. Free for Chrome.
Get TabRollbackImportant limitation
This reads an existing PDF text layer; it does not perform OCR. Scanned pages that contain only images need a separate optical character recognition tool. Each line on the page becomes its own line in the TXT file, a blank line separates one page from the next, and a word split by a hyphen at the end of a line stays split.
Paragraphs, tables and footnotes in the TXT file
Lines are rebuilt from the height at which each piece of text sits on the page. Blank lines inside a page are not kept, so paragraphs follow one another line by line; add breaks after pasting if you need them. A table row comes out as a single line with spaces between the cells, not tabs, so pasting it into a spreadsheet puts the whole row in one cell. A footnote marker or subscript raised or lowered by more than a couple of points can land on a line of its own.
Keep headings on when some pages may be empty
With page headings turned off, a page that has no text is skipped entirely and leaves no gap, so nothing in the TXT file shows where it was. Leave the box ticked when the PDF mixes scanned and typed pages, or when you will need to cite page numbers later. Headings use positions in the file, page 1 being the first page of the PDF, not the number printed in the footer.
Counting words for a quote, an essay or a translation
The result line reports words and characters for the whole TXT file. Words are counted between spaces and line breaks, so each heading line such as --- Page 3 --- adds four, and the character count includes line breaks. For a count of the body text alone, untick the headings box before you extract.
Extract selectable text from a PDF
Use this tool when a PDF already contains a text layer and you need an editable plain-text copy. Reports, exported documents, ebooks, invoices and digitally created forms often include selectable text even when the page design is complex. PDF North reads that text inside the browser, preserves the selected page order and creates a UTF-8 TXT file that can be opened in a text editor, notes app or writing tool.
Extract every page or choose a page range
Leave the page field blank to process the complete document. To work with one section, enter individual page numbers or ranges such as 1-3, 7, 10-12. Invalid or out-of-range selections are rejected before a file is created. Optional page headings make it clear where each PDF page begins in the text output, which is useful when reviewing a long report or quoting material with page references.
Preview, copy and download the result
After extraction, the result panel shows a text preview plus the number of processed pages, words and characters. The copy button places the complete result on the clipboard, including text beyond the preview limit. The TXT download contains the complete extraction and uses Unicode encoding so common accented characters and symbols are retained.
Why text order can differ from the page
A PDF stores text as positioned drawing instructions rather than as a normal flowing document. Two columns, tables, floating labels, headers and individually positioned characters can therefore appear in an unexpected reading order when converted to plain text. PDF North follows the text items supplied by the PDF engine and separates visible lines, but it does not recreate columns, tables, fonts, images or the original layout. Review the preview before relying on the output.
Scanned PDFs need OCR
This tool does not run optical character recognition. A scanned page may look like text while actually containing only a photograph of the page. If the selected pages have no text layer, PDF North stops with a clear message instead of creating an empty or misleading download. Use a dedicated OCR tool for scans, handwriting or photographed documents, then check the recognised text carefully because OCR can misread names, numbers and punctuation.
Private PDF-to-text conversion
The PDF is opened with Mozilla PDF.js and processed in this browser tab. It is not uploaded to PDF North or stored in an account. Temporary download URLs are released when the page closes or reloads. Password-protected PDFs are not supported, and the 60 MB and 250-page limits reduce the chance of exhausting device memory.
Useful next steps
If you only need part of the document as another PDF, use Extract PDF Pages instead so the original layout remains intact. Use PDF to JPG when you need page images rather than editable words. For documents being shared outside their original workflow, the PDF metadata remover can inspect and remove common hidden document properties after you finish other edits.
Frequently asked questions
Will text extraction read a scanned page?
No. The tool reads an existing text layer only. If every selected page contains only images of text, it stops instead of inventing words. When only some pages are scans, it still runs: text pages are extracted and each scanned page is left empty, or shows just its --- Page 5 --- heading. A scan that already went through OCR elsewhere, often called a searchable PDF, carries an invisible text layer, and that layer is read. PDF North does not run optical character recognition.
Can I extract text from only part of the PDF?
Yes. Leave the page field blank for every page, or enter ranges such as 1-3, 7. Pages come out in the order you type them, so 9, 1-2 puts page 9 first, and a page listed twice is extracted once. The headings box is ticked by default and puts a line such as --- Page 3 --- above each page; untick it for clean text to paste. The preview stops at 20,000 characters, but Copy all text and the TXT download include the full result.
Why does the extracted text order differ from the page?
A PDF stores text as positioned drawing instructions. Columns, tables and floating labels can come out in an unexpected order. Pieces of text at the same height are joined into one line with a space, so a label and its value can stay together, while two columns side by side may be interleaved. This tool does not rebuild the layout, fonts or images. Read the preview before you rely on the TXT file.
Is the PDF uploaded to extract the text?
No. The text is read from the file in this browser tab. The TXT download is created locally and is not sent to PDF North. A file named report.pdf is saved as report-text.txt in UTF-8. The word and character counts include the heading lines when headings are on. Password-protected PDFs are not supported, and the limits are 60 MB and 250 pages.
Why does the extracted text show odd symbols or wrong letters?
Some PDFs draw text with fonts that do not record which letter each shape stands for. The page looks normal, but the text layer holds codes that come out as boxes, symbols or the wrong letters. A plain-text export cannot repair that mapping. If copying from a PDF reader gives the same garbled result, the problem is in the file, and OCR on page images is the practical route. PDF to PNG creates those images on this device.
Will hidden or white text be extracted too?
It can be. The tool reads every text item stored on the page, not only what is visible. White text on a white background, words covered by a picture and the invisible layer that OCR software places behind a scan can all appear in the TXT file. Read the output before you quote or publish it, and do not assume that words hidden under a black box are gone: covering text is not redaction.
Does the TXT file keep bold text, links or images?
No. A TXT file holds characters and line breaks only. Bold and italic styling, font sizes, colours, images and the web addresses behind links are not written to it. The visible words of a link are extracted when they are drawn on the page, but the address itself appears only if it is also printed there.
Related guides
These how-tos use the same local tool. The full list lives on PDF guides.
- Extract Text from a Scanned PDF: What Local Tools Can Do
- PDF to Word: Get Editable Text Without a Fake DOCX