About this tool
Turn a searchable PDF text layer into reviewable plain text with explicit page selection, ordering policy, empty-page reporting, and browser-safety limits.
PDF to Text Converter reads the text layer of a PDF with PDF.js and writes it out as plain UTF-8 text, page by page. It works on documents that already contain selectable text, such as exported reports, invoices, contracts, and e-books, and it tells you plainly when a page has no text layer at all instead of returning silence. You choose which pages to extract, whether to follow the PDF's internal content order or regroup text into visual lines, and whether to insert a separator before each page. Alongside the text you get word and character counts per page and a JSON manifest that records the settings and the empty pages. Files up to 50 MB and 10,000 pages can be opened.
- Uses PDF.js to read the real text layer instead of guessing from raw PDF bytes.
- Supports up to 500 selected pages with bounded range parsing, optional page separators, and content-flow or visual-line ordering.
- Reports page-level words, characters, and empty text layers, then exports UTF-8 TXT and JSON evidence without performing OCR.
How to use PDF to Text
Drag a PDF onto the drop area or press Choose PDF file. Once the signature check passes, the page count appears. In the Pages field, type a selection like 1-3, 5, 8; leaving it blank extracts everything, but only when the document has 500 pages or fewer. Pick a Text order: Content flow follows the order text items were written into the PDF, while Visual lines groups nearby baselines and sorts them left to right, which usually reads better for single-column pages. Tick Include page separators if you want a marker before each page. Press Extract selected pages, watch the progress bar, and use Cancel extraction if it runs too long. Then copy from the output box, press Download TXT, or save the Manifest.
When this tool is useful
- Pulling quotes from a long PDF report into notes without retyping or fighting with column-aware copy and paste.
- Legal or compliance staff who need the text of specific contract pages for a search-friendly archive.
- Developers feeding document text into a search index, embeddings pipeline, or diff tool that expects plain text.
- Students extracting selected chapters from a textbook PDF for annotation in a plain-text editor.
- Checking whether a scanned PDF actually has a text layer before deciding whether OCR is needed.
Practical tips
- If every page reports as empty, the PDF is an image scan. This tool does no OCR, so run an OCR step first and come back.
- Try both text orders on one tricky page. Content flow can interleave columns; Visual lines fixes that but may merge side-by-side table cells.
- For documents over 500 pages, extract in chunks such as 1-500 and 501-1000 so each run stays under the selection limit.
- Output caps at 5,000,000 characters and 1,000,000 per page; dense legal documents can hit this, so narrow the range when truncation is reported.
- Keep the JSON manifest with the text. It records per-page word counts and which pages were empty, which helps when someone questions a missing paragraph.
Examples you can test
Load an example, compare the result with the expected output, then replace it with your own input.
Selected pages of a report
Example input
quarterly-report.pdf, 48 pages, Pages field set to 2-4, 9, Visual lines, separators on
Expected output
Four page blocks in the text box, each preceded by a separator, with totals like 2,310 words · 14,880 characters · 0 empty pages.
Page 9 keeps its original number in the separator even though it is the fourth block in the output.
Scanned document
Example input
scan.pdf, 12 pages of photographed paper with no text layer
Expected output
Extraction completes quickly with 0 words and 12 empty pages; the text area stays blank.
An empty result here is a finding, not a failure: the file needs OCR before its text can be extracted.
Validation checklist
- Confirm the page count and empty-page count in the totals line match expectations.
- Spot-check numbers, names, and headings on at least one page against the PDF.
- Review reading order on any page with columns, tables, or sidebars.
- Make sure the Pages selection did not exceed 500 pages, or split the run.
- Download the manifest if the extraction will be cited or audited later.