To extract text from a PDF online for free, open it in the PDF to Text tool: its text is extracted page by page, in reading order, ready to copy or download as a .txt file. Everything runs in your browser and nothing is uploaded. One condition decides whether it works at all: the PDF must contain a text layer. A scanned PDF is images, not text, so it contains nothing to extract — that job needs OCR instead.
Text Layer Versus Scan: The Deciding Test
Before extracting, test the file in any PDF reader: try to select a word with the mouse. If the word highlights, the file has a text layer and extraction will work. If nothing highlights, the pages are pictures of text, and no extractor can pull words out of them — the tool would return empty pages, correctly, because there is no text stored in the file. This single test saves more time than any other step in the workflow, because it separates the two jobs cleanly: text-layer PDFs go to PDF to Text, scanned PDFs go to OCR.
Even with a text layer, extraction is plain text, not layout. Fonts, columns and exact positioning are not reproduced; what comes out is the file's own characters, grouped into lines by their positions, in the order the file stores them. For normally produced PDFs that is reading order. For PDFs assembled from odd sources, the stored order can be strange — a two-column page may interleave lines from both columns — and no extractor can fully undo that. The extracted text is therefore a working copy for notes, quotes and search, not a replica of the page design.
Using the PDF to Text Tool, Step by Step
- Choose or drop your PDF. Extraction starts immediately and the status line counts through the pages, so you can see progress on a long document.
- Read the extracted text in the box. It is editable, so you can trim boilerplate — headers, footers, page numbers — before saving. Page markers show where each page's text begins, and the counter shows characters and words.
- Copy the text, or download it as a .txt file named after your PDF. The .txt opens anywhere: notes apps, word processors, email drafts, search indexes.
- Check quotes against the original. When a passage will be cited, compare it with the PDF page it came from. Page markers make this easy: a quote from page 23 should be checked on page 23, not trusted from the extracted copy alone.
Tip: Trim headers and footers before you download, not after. Repeated running heads — a report title on every page, a footer with a page number — pollute word counts and search results in the saved text, and they are far easier to remove in the extraction box, page by page, than from a finished .txt file.
Extraction Versus OCR: Which Tool for Which File
The two text-recovery jobs look identical from the outside — a PDF goes in, text comes out — but they work on different inputs and fail differently. The table below keeps them apart, because using the wrong one is the single most common failure in this workflow: an extraction that returns empty pages is almost always a scanned file, and an OCR run on a text-layer file is wasted effort.
| Your file | Right tool | What comes out | Failure sign |
|---|---|---|---|
| Text-layer PDF (words select) | PDF to Text | Plain text, page markers, editable | Odd line order on multi-column pages |
| Scanned PDF (words do not select) | OCR tool | Recognised text, new text layer | Recognition errors on faint print |
| Mixed PDF (some pages scanned) | Extract, then OCR the gaps | Text for typed pages, OCR for scans | Empty pages in the extraction |
| Image-only PDF (photos) | OCR tool | Recognised text from images | No text layer to extract |
| Password-locked PDF | Unlock first, then extract | Plain text from the unlocked copy | Extraction blocked until unlocked |
Worked Example: Extracting a 48-Page Council Report for Notes
A researcher receives a 48-page council report as PDF only and needs its findings in her notes app. One extraction gives her the full text with page markers, so a quote can still be cited by page — and an hour of retyping becomes a minute of tidying. The report's structure shows why page markers matter: the introduction runs 4 pages, the findings run 26 pages, and the appendices run 18 pages, totalling 4 + 26 + 18 = 48 pages. If the extraction box shows a different total, a section was missed or a page range was wrong, and the fix is to re-extract from the original rather than patch the text file. After trimming the repeated running heads, the notes file holds the findings in order, each passage traceable to its page.
| Section of the report | Pages | Running total | Use in notes |
|---|---|---|---|
| Introduction | 4 | 4 | Background and scope |
| Findings | 26 | 30 | Quotable evidence, cited by page |
| Appendices | 18 | 48 | Tables and references |
| Whole report | 48 | 48 | Full text, page-marked |
Pages per section in the 48-page council report example — hover or tap each bar for details.
Getting the Best Result When You Extract
Extract in sections for long documents, not all at once. A 200-page report extracted in one pass produces a text file that is hard to navigate and harder to clean: running heads repeat two hundred times, and a single mis-ordered section pollutes the whole file. Extracting by section — introduction, findings, appendices — gives you smaller files with their own page markers, each easy to check and tidy before moving on. The total work is the same; the errors are contained.
Build a small cleanup routine for every extraction. Remove running heads and footers first, then fix the predictable artefacts: hyphenated words split across line breaks, page numbers standing alone on a line, and table rows that collapsed into single lines. Ten minutes of cleanup turns raw extraction into notes you can actually quote from, while skipping it leaves a file that looks complete but misleads on detail — a hyphenated word rejoined wrongly changes a term, and a collapsed table row changes a figure's meaning.
Keep the PDF beside the text file, not instead of it. The text file is for searching, quoting and drafting; the PDF remains the authority for layout, figures and signatures. When a quote matters — a contract clause, a reported statistic, a cited finding — verify it against the PDF page the marker points to, and record the page number with the quote. A notes file without page references is a draft, not evidence.
Common Mistakes When Extracting PDF Text
1. Extracting a scanned file and concluding the tool is broken. Empty output from a scan is correct behaviour: there is no text layer to extract. Test selectability first, and send scans to OCR.
2. Keeping running heads in the saved text. Repeated headers and footers distort word counts and clutter search. Trim them in the extraction box before downloading.
3. Trusting multi-column order blindly. Two-column pages can interleave lines from both columns in the stored order. Read extracted passages against the original page before quoting them.
4. Losing the page reference. The page markers are the citation system. Stripping them before saving turns quotable evidence into unsourced text; keep them, or record the page number alongside each quote.
5. Editing the PDF instead of the text. Extraction produces a working copy for notes and quotes. If the PDF itself needs changing, that is an editing job for a PDF editor, not an extraction job.
Key takeaway: Extracting text from a PDF recovers the file's own text layer, page by page, as editable plain text — fast and faithful for text-layer files, empty by design for scans. Test selectability first, trim running heads before saving, and keep page markers so every quote stays citable.
Frequently Asked Questions
How do I extract text from a PDF online for free?
Open the PDF in the PDF to Text tool: text is extracted page by page, in order, ready to copy or download as a .txt file, all in your browser.
Why is the extracted text empty?
Almost always because the PDF is scanned images with no text layer. Open it and try to select a word with your mouse — if you cannot, there is no text to extract, and OCR is the right tool.
Does it keep the page structure?
Yes — every page starts with a page marker, and lines follow the page's own line breaks. Fonts, columns and exact layout are not reproduced; this is plain text.
Can I edit the text before downloading?
Yes, the output box is a normal text area. Trim headers, footers or whole pages, then download or copy what remains.
What order does the text come out in?
The order the PDF itself stores it, grouped into visual lines — which for normally produced PDFs is reading order. PDFs assembled from odd sources can store text strangely, and no extractor can fully undo that.
Is my PDF uploaded?
No. Extraction runs in your browser on your device; the document never leaves it.
What is the difference between PDF to Text and OCR?
PDF to Text reads the file's existing text layer. OCR recognises text from images. Use extraction for text-layer PDFs and OCR for scanned PDFs; the selectability test tells you which one you have.
Need a website that handles documents, downloads or enquiries cleanly? If your business needs clear tool pages, pricing and a contact flow that converts visitors into enquiries, Talk to us about your website.
For the surrounding document workflow, pair extraction with Organize PDF for page order and Merge PDF for combining files. More step-by-step tool guides live on Asad Webs blog.
Be the first to share a useful question or thought.