Validation errors in a Kindle-converted Epub

You converted an Azw3 or Mobi to Epub, ran it through a validator, and got a wall of errors. The good news: most of them are Amazon leftovers, they follow a pattern, and none of them touch your text.

Amazon's formats are built from HTML plus Amazon's own extensions: page number maps, bookkeeping attributes, markers their tools use internally. When a book gets converted to Epub, much of that residue comes along for the ride. It means nothing to an Epub reader, most readers silently ignore it, but a validator flags every occurrence, and stores reject the file. Here is what each leftover looks like and how to clear it.

page-map on the spine

Reported as RSC-005: attribute "page-map" not allowed here on the OPF. This is Amazon's old page numbering extension. Delete the page-map attribute from the <spine> element. If the manifest lists a page-map.xml entry, you can remove that too, along with the file.

pageList blocks in toc.ncx

Reported as RSC-005: element "pageList" missing required attribute or similar complaints about pageTarget. Same page numbering system, different file. Converters write these blocks incomplete. If you want print page numbers you can fill in the required attributes, but for a converted novel the simpler move is deleting the entire <pageList>...</pageList> block. Regular navigation lives in navMap and is unaffected.

data-AmznRemoved and friends

Attributes like data-AmznRemoved, data-AmznRemoved-M8 and amzn-src-id are markers Amazon's tooling left in the markup. They carry no meaning outside Amazon's pipeline. Delete them. They tend to appear hundreds of times, so use search and replace in an Epub editor rather than fixing hits one by one.

Epub 3 vocabulary in an Epub 2 file

Reported as RSC-005: attribute "epub:type" not allowed here, with role and aria-* attributes in the same boat. The converter wrote Epub 3 markup into a book packaged as Epub 2. Deleting the attributes is the quick fix and costs you nothing in an Epub 2 book. (Upgrading the whole package to Epub 3 also resolves it, but that is bigger surgery than this cleanup calls for.)

Finding every occurrence

The checker groups all of this by problem rather than by line, tells you which group is which, and shows the source for every hit. Run the book through it once before you edit, so you know the full list, and once after, to confirm the report is clean.

Check your converted Epub

Converting a book here? The Mobi to Epub converter and the checker live on the same site, so the round trip is short. Both run in your browser and your books never get uploaded.