PDF to Word (DOCX)
Best-effort pdf to word (docx), from extracted text and a heading/list heuristic.
Runs on your device. The file is never uploaded.
PDF to Word (DOCX) extracts the text with pdf.js, turns it into Markdown, then hands that to pandoc's docx writer, which returns the .docx as bytes. The output is your file's stem plus .docx. Nothing but the words crosses over: pictures, typefaces and colour are left behind, because everything passes through plain Markdown first.
Questions
Does my PDF get uploaded to convert it?
No. The whole conversion runs in a Web Worker in this tab. pdf.js reads the text out of your PDF, then pandoc compiled to WebAssembly writes the .docx. Nothing is sent to a server, there are no accounts, and the page keeps working with your network switched off after the first visit.
Will the Word file look like my PDF?
No, and it does not try to. You get the text with a heading and list guess laid over it. The most common glyph height in the document is treated as body text, any line more than 15 percent taller becomes a heading, and a line starting with a dash, an asterisk or a bullet becomes a list item. Columns, tables, page breaks and exact positioning are not reconstructed.
Why does it say "no extractable text found in this PDF"?
Because pdf.js found no text at all. That is normal for a scan or a photo of a page, where the file holds an image rather than characters. Run PDF OCR first to add a text layer, then convert the result. The same message appears for a PDF whose pages are genuinely empty.
Are images, fonts and colours carried into the DOCX?
No. Only text crosses over. Embedded pictures, logos, the typefaces you were looking at and any colour are all dropped, because the conversion goes through plain Markdown before pandoc writes the Word file. To get the pictures, run Extract PDF images. To keep the page exactly as it looked, render it with PDF to JPG or PNG instead.
What is the output file called?
Your file name with .docx in place of .pdf, so report.pdf comes back as report.docx. One PDF goes in and one document comes out; your original is never modified. The .docx is written through pandoc's output-file path rather than through standard output, because a Word file is binary and cannot survive being carried as a text string.
How is this different from PDF to Markdown?
They share the first half. Both run the same pdf.js extraction and the same heading heuristic. PDF to Markdown stops there and hands you the .md. This tool passes that same Markdown to pandoc with the docx writer. So the Word file can never hold more than the Markdown did, and if the Markdown comes out wrong, the DOCX is wrong in exactly the same way.