Bringing a book out of Google Docs, Pages or Scrivener
Almost nobody writes in the format a typesetter wants, and that is fine — the writing tool is for writing. The export is a translation, and like any translation it can be careless or careful.
What makes it worth ten minutes is this: the typesetter never sees your Google Doc, your Scrivener project or your Pages file. It sees the one file you exported. Whatever got lost, mangled or invented on the way out is now the book.
Here is what each tool does badly, and what to do about it.
Google Docs
Export as .docx, not PDF. File → Download → Microsoft Word (.docx). A PDF is a picture of pages; it has no chapters, no paragraphs, no italics a machine can find. Downloading a PDF because it looks finished is the single most expensive mistake in this guide.
Watch for headings that are not headings. Google Docs makes it easy to produce a chapter title that is bold, larger and centered and still carries the paragraph style Normal text. It looks correct and says nothing. Put the cursor in each chapter title, check the style dropdown reads Heading 1, and fix the ones that do not.
Check the quotation marks. Smart quotes are on by default but pasted text arrives with whatever it had. Search the file for a straight " and a straight ' after you export and see what turns up.
Anything pasted from the web brings its own clothes. A paragraph copied out of a browser carries fonts, sizes, colors and a background you cannot see against white, and it arrives in the export as direct formatting over your styles. Direct formatting wins. Repair it by selecting all, clearing formatting, and re-applying your styles.
Pages
Export as .docx. Share → Export To → Word. A .pages file is readable by Pages and nothing else, including every typesetting tool in existence.
Text boxes and floating images do not survive the trip. Pages is half word processor and half page-layout program, and the page-layout half has no equivalent in a flowed book. A pull quote in a floating box, a photograph beside a paragraph, a rule anchored to the page — a book page is one column of set type, and these have nowhere to go. Move what you need into the body text as its own paragraph before you export, and delete the rest.
Confirm the section breaks came through as breaks. Pages sections carry page setup, headers and layout, and on export some become plain paragraph breaks. Open the .docx and look at where each chapter starts.
Scrivener
Compile, do not export. Export dumps the text; Compile is the feature that builds a document, and it is the one to use. Choose a plain manuscript format rather than one of the formats that bakes in a design — you want the words and their structure, not Scrivener's idea of a paperback.
Turn off the extras. Compile will happily insert separators between documents, a title page, a copyright page, a table of contents and a page break before every folder. Turn all of it off. Your typesetter builds front matter and scene breaks itself, and a duplicate title page is worse than none: it is a page you have to find and delete later.
Check the section types. Scrivener decides what becomes a chapter heading from how each document's section type is assigned, not from what the text looks like. Get that right in Compile and the chapters arrive as chapters.
In the Interior Formatter. It reads the .docx you exported and keeps only the structure — chapter starts, italics, bold, scene breaks — then sets the book itself. It finds chapter starts even in files with no styles at all, so a Google Docs export full of bold 16-point titles still comes through. Start a book → · about the Interior Formatter
Word
You are already in the native format, so there is no translation to get wrong. There are two things to do anyway.
Accept or reject every tracked change, then delete every comment. A file with revisions still live can import either version of a sentence, and the one it picks is not the one you meant.
Save a copy. Work on the copy. The cleanup below removes things, and you want the original intact.
The cleanup that matters more than the source
Whatever tool you came from, the same short pass fixes most of what is left.
- Chapter starts are real headings. Heading 1, applied to every one of them, with none missed.
- One body style for all prose. Not three near-identical styles built at different times.
- No manual line breaks inside paragraphs. Shift+Enter looks like a new line and is not a new paragraph. Search for them and take them out.
- No empty paragraphs used as spacing. Four returns to push a chapter down the page is four empty paragraphs, and the typesetter sets that space itself.
- Save as .docx. Not PDF, not RTF, not
.pages, not a Scrivener project.
All of it is covered in more depth in Preparing a manuscript that typesets cleanly, which is worth reading before you export rather than after.
In the Interior Formatter. Scene breaks come through as breaks rather than as blank lines, and you pick the mark that prints in their place. Previews are free and unlimited, so you can upload an export, look at the result, fix the manuscript and upload again as many times as it takes. Start a book → · about the Interior Formatter
Open the file before you upload it
Find the .docx you just exported. Open it — in Word, in Google Docs, in anything that reads it — and scroll the whole book.
You are looking for four things: every chapter starting where it should, no stray title pages or separators the compiler added, italics still italic, and no boxes or images sitting somewhere strange. Five minutes, and it catches the errors that are otherwise invisible until the book is set.
That file is the book now. The original is just where you wrote it.
Make yours
Everything here is built into the tools, and they are free to use: name your book, bring your manuscript or your artwork, and work as long as you like. You only pay when you want the files without the preview watermark.
Next: Preparing a manuscript that typesets cleanly · Also: Chapter openings and scene breaks · Images inside a book · How to format a book interior