AZW3 to PDF Converter turns an AZW3 Kindle ebook into a PDF you can print, annotate or open on any device without a Kindle app. AZW3 is the KF8 Kindle format: modern XHTML and CSS cut into skeleton and fragment records that only Kindle software knows how to stitch back together. The converter reassembles those records into chapters, restores the links and images, and then sets the book on fixed pages in the size you choose: A4, US Letter, A5 or a 6 by 9 inch book page, with adjustable text size and margins. The PDF contains real text rather than pictures of pages, so it stays small and you can select, copy and search it. The cover becomes the first page, a contents page links to every chapter, and PDF bookmarks follow the Kindle table of contents. The embedded font covers Latin with all European diacritics, Cyrillic and Greek. Older .azw, .mobi and .prc files are accepted too. Everything runs in your browser, so the book is never uploaded. AZW3 files from the Kindle store are DRM protected and cannot be converted by any tool without removing that protection, which this tool does not do. To keep a reflowable ebook instead use AZW3 to EPUB Converter, and to read the book on screen use MOBI Reader, which opens AZW3 files as well.
An AZW3 file looks like a single ebook, but inside it is a Palm database: a header followed by numbered records. The first record holds the MOBI header and the EXTH block with the title, author, publisher and cover reference. The text records come next, compressed in blocks of 4096 bytes, followed by the images, one per record. What makes KF8 different from the older MOBI format is how the text is stored. Instead of one long HTML document, the book is split into skeletons, which are the outer shells of each chapter file, and fragments, which are the pieces of content that get inserted into those shells at recorded positions. Two index tables describe where every skeleton and fragment begins, and a third maps the original file names and CSS. Links point to a fragment number and an offset inside it rather than to a named anchor. To make a PDF, the converter has to undo all of this: decompress the text, rebuild each chapter by inserting fragments into their skeletons, translate fragment links into real anchors, match image records to the pictures that reference them, and read the navigation index to recover chapter titles. Only then can the book be laid out on pages. Many files saved as .mobi by Calibre or Kindlegen actually contain both an old MOBI version and a KF8 version; in that case the KF8 part is used, because it keeps more of the original structure.