pdf-to-docx

PDF to Word

Turn a PDF into a Word document with real paragraphs, headings and tables.

Reads: a PDF — application/pdf
Writes: one file — Word documents
Tool page: /pdf-to-word

Example

Hand a document to somebody who works in Word.

Given

report.pdf — a PDF of written text

From code

            import { loadDocument, runOperation } from '@pagecraft/core';

const report = await loadDocument({ bytes, name: 'report.pdf' });

const [output] = await runOperation('pdf-to-docx', [report], {
  headings: true,
  tables: true,
});

From a shell

            pagecraft pdf-to-docx report.pdf

What comes back

One Word document.

This example is run by the test suite. If it stopped being true, the build would fail.

Options

`pages--pages`

string, a page selection: '1-3, 7, 9-', 'even', 'last'

Pages

Which pages to convert.

Default: 'all'

`headings--headings`

boolean

Find headings

Treat larger and bolder lines as headings.

Default: true

`tables--tables`

boolean

Find tables

Turn ruled and aligned columns into Word tables.

Default: true

`images--images`

boolean

Keep pictures

Carry the pictures across, at the resolution the PDF stored them.

Default: true

`pageBreaks--page-breaks`

boolean

Break at page ends

Start a new page in Word wherever the PDF started one.

Default: true

What it cannot do

This is a conversion, not a round trip. It gives back an editable document that says what the PDF said; it does not give back the page, and the two will not look the same side by side.
A PDF does not record what was a heading, a list or a table; it records type at sizes in places. All of that is inferred from the page, and a page laid out unusually will be read unusually.
The faces, sizes, colours and exact positions of the original are not carried across, and neither is anything drawn rather than typed: a chart, a logo outline, a signature.
A picture comes across at the resolution the PDF stored and the size the page drew it, in the place a reader would have met it — but a picture the text wrapped around lands after that text rather than beside it.
Scans stored as CCITT fax, JBIG2 or JPEG 2000 are left out: Word cannot read any of them.
A scanned page holds a picture of words. Run OCR over it first, or the document will come back with the pictures and none of the text.