Convert PDF to Word

Yes. This beta reads the PDF text layer with PDF.js and generates a .docx on your device. The site must first load its processing code, but your selected document is not sent to a conversion server or saved in persistent browser storage. Keep the original PDF; the Word file is a separate download.

Yes. The online text-only beta is free to use, with no account or watermark. It creates editable paragraphs from PDFs with selectable text, but it does not run OCR itself or preserve images, table structure, columns or the original page layout. OCR a scan first if needed, then review the DOCX beside the source PDF.

Yes. Extracted text becomes real Word paragraphs, not images of PDF pages. You can select and edit it in Word or LibreOffice. PDF text order is not always the intended reading order, so check the result before relying on it.

No. This is a text-only beta for simple documents, not a layout replica. It uses a standard font and reconstructs basic paragraphs. Images, table structure, columns, original styling, interactive forms and annotations are not reproduced. Text in tables or columns may be flattened into an unsuitable reading order and need manual editing.

Not directly in this beta. A scan without a usable text layer is refused rather than producing an empty Word file. You can use the separate browser-local English OCR tool, which supports up to 20 pages, then select its searchable PDF here. Proofread OCR output: recognition errors will carry into the Word file.

If no page contains usable text, no DOCX is created. In a mixed document, pages without text are left blank, their source page numbers are listed, and you must acknowledge the warning before downloading. They may be scans, images or intentionally blank pages; this tool cannot reliably tell those apart.

Not necessarily. A break is inserted between each source page, including pages left blank, but Word reflows paragraphs with its own fonts, margins and line breaking. One source page may take more than one Word page. Check the result in your editor rather than treating the source-page count as an output-page count.

The beta accepts one PDF up to 25 MB and 100 source pages, with a maximum of one million extracted text characters. It refuses larger documents rather than exporting a silent prefix. Password-protected and permission-locked PDFs are not supported. Available device memory can impose a lower practical limit. No account is required and no watermark is added.

The original PDF is not changed, but the new DOCX does not carry its digital signatures or verification status. This is not a redaction or sanitization tool: hidden text in the PDF text layer can become visible in Word, even if it was covered by a black rectangle. Review sensitive output carefully and keep signed originals.