Out of a document

PDF to EPUB

Reflowable text for an e-reader.

On this device

Convert to EPUB

The sections it found are listed before you download, so you can check them first.

    Opened on this device · never transmitted

    What this tool does, exactly

    A PDF stores a page; an e-reader has a small screen and a reader who has already chosen their type size. This converts the one into the other, so the text reflows instead of being shrunk to fit.

    What it does
    Reads the text a PDF stores and rebuilds a document from it. Type size becomes heading levels, bullets and indentation become lists, and headings at the top two levels become chapters with a working table of contents.
    What the PDF must contain
    A text layer, since a scan has nothing to reflow. The tool names the pages instead of returning a book of empty chapters, and OCR a PDF is the tool that reads a scan.
    What it does not carry
    The layout, and deliberately. Columns, exact spacing, fonts and the source's page breaks are dropped, because an e-reader paginates to its own screen and a reader has chosen their own type size. An EPUB that fought those choices would be a PDF with extra steps.

    About this tool

    Why not just put the PDF on my e-reader?

    You can, and that is why the problem is familiar: a PDF has a fixed page width and fixed type size, so on a six-inch screen the whole page is shrunk to a postcard and you pan around it. EPUB stores the text rather than the page, so the reader lays it out at the size you chose.

    Will the EPUB work on a Kindle?

    Yes. Amazon stopped accepting MOBI for Send to Kindle in 2022 and takes EPUB directly now, so EPUB is the current format for both Kindle and Kobo. Kobo has always read it natively.

    My PDF came back as “no text layer”. Why?

    Because it is a scan, a picture of text rather than text. There is nothing to reflow, and the tool says so and names the pages rather than handing you a book of empty chapters.

    How good are the chapters it finds?

    They are inferred, because a PDF never records that something is a chapter. Headings at the top two levels become chapters, and a heading with nothing above it is ignored, otherwise a running header at the top of every page would give a table of contents with one entry per page.

    Does my document leave the device?

    No. The text extraction, the structure inference and the packaging all happen in this tab.

    More out of a document tools · All 55 · How they work