PDF to Word

Four conversion modes — pick the one that fits your document:

• Layout (default) — uses pdf2docx to preserve tables, columns, and figure positions. Best for forms, reports, invoices.
• Smart structure — analyses font sizes to detect headings, lists, and paragraphs, and emits a Word doc with proper heading styles (visible in Word's navigation pane). Best for articles, blog posts, books, and documentation. Drops tables and figures.
• Flowing text — extracts text in reading order, emits one paragraph per block. No structure detection. Always produces clean output even on awkward PDFs.
• Marker (ML) — uses an ML model for structure understanding. Best fidelity for academic papers, books, and complex documents. Not installed — run pip install marker-pdf. First run downloads ~2 GB of models. Conversion is slow on CPU (30–60s/page).

If your PDF is a scan, run it through OCR PDF first.