PDF to Word, with the limits stated up front
Convert the text layer of a PDF into an editable .docx. Text-led documents come through well; multi-column layouts, text boxes and complex tables need tidying afterwards — and scans need OCR first.
Why PDF to Word is genuinely hard
PDF is a print format: it says “draw this glyph at this coordinate”. Word is a flow format: it says “this is a paragraph, reflow it when the page changes”. Conversion is therefore not a translation but an inference — the tool has to guess which glyphs form a sentence, which sentences form a paragraph, and which paragraphs belong under a heading.
When the source PDF was exported from a well-structured Word file, the guess is usually right. When it was designed in InDesign, assembled from scans, or laid out with absolute positioning, the guess is often wrong, and no converter — including paid ones — can recover structure that was never in the file.
What converts well, and what does not
- Good: reports, articles, letters, contracts and manuals with a single text column and real text.
- Good: headings and bulleted lists, which are detected from size and indentation.
- Poor: two- and three-column academic papers, brochures and newsletters.
- Poor: forms, invoices and anything built on positioned text boxes.
- Impossible: scanned pages, which have no text layer at all.
How to convert
- Add the PDFThe preview reports whether the file has a text layer — the single best predictor of quality.
- Choose flowing text or page breaksFlowing text is better for editing; page breaks preserve the original pagination at the cost of odd gaps.
- Export and fixOpen the .docx, set the styles you want, and expect ten minutes of tidying on anything with tables.
| Item | Behaviour |
|---|---|
| Input | Text-based PDF (no scans, no OCR) |
| Output | .docx with editable paragraphs and headings |
| Layout | Simplified — columns and positions are not reproduced |
| Images | Not carried over — this build writes text only |
| Tables | Rebuilt as text lines, not as Word tables |
FAQ
Will my formatting be preserved?
Font sizes and bold/italic are approximated; exact fonts, colours, spacing and page layout are not. Treat the .docx as a good first draft that you restyle, not a replica.
Does it work on scanned PDFs?
No. A scan is an image of a page. Without OCR there is nothing to convert; the output would be empty. OCR is a separate step and its results always need proofreading.
Is this better than copy-pasting from the viewer?
For long documents, yes — one pass instead of page-by-page copying, and paragraph breaks are preserved. For a single page, copying from the viewer is often faster.