How to catalogue a large ebook collection by author, series and subject

Most ebook collections were never planned. They grew: a bundle here, a course reader there, a folder of PDFs from a degree, a decade of purchases from a shop that no longer exists. The result is a directory of files called things like book(3).epub, and the question of what you own is answered by scrolling.

Cataloguing it does not mean typing it all in. Most of what a catalogue needs is already inside the files.

Start from what the books say about themselves

An EPUB carries a small record of itself: its title, its authors, often its publisher, language, identifiers such as an ISBN, subjects, and sometimes the series it belongs to and its place in it. That record was written by the publisher, which makes it the best source you have. A PDF carries less, and what it carries is often wrong (a PDF "titled" Microsoft Word - draft7.docx is common), but its first pages usually say what it is.

So the first rule: read the metadata from the files, and trust it over filenames. A filename was typed by whoever saved the file; the metadata was set by whoever made the book.

Decide what counts as "a book"

A large collection soon holds the same novel twice: an EPUB and a PDF, or two translations, or the 2003 and 2019 editions of a textbook. A catalogue that lists files shows you three Middlemarches. A catalogue that groups them shows you one work with its editions, and lets you pick which to open.

Series and their order

Series are where file-based collections fall apart first, because the order lives in no filename. EPUBs often record the series and the book's number in it; when they do, use it. Make sure your tool reads that field rather than ignoring it, because "what comes next" is one of the questions a library is for.

Subjects you can browse

Publishers tag books with subjects. Those tags are uneven, but taken across a whole collection they give it a shape: a list of subjects with a count beside each, which tells you at a glance that you own forty books on statistics and two on Byzantium. That is the view a pile of files never gives you.

Duplicates, and what to do with them

Two kinds turn up. The same file twice (identical bytes, saved twice) should simply be recognised and kept once. The same work in two files is not a duplicate to delete; it is two editions of one work, to be grouped.

Gaps: leave them honest

Some files have no usable title at all. The temptation is to fill the gap with the filename, but a guess recorded as fact will outrank the truth forever. Better to leave the gap visible, or to let something read the book's title page and suggest a title, marked as a suggestion until someone accepts it.

The tools

Calibre is the long-standing answer, and for good reason: a free desktop application that reads and edits metadata, converts formats and sends books to e-readers. If you want to curate every field by hand on one computer, it is hard to beat. Calibre-Web puts a browser front-end on a Calibre library, so you can browse and read it from other devices.

Lexarii takes the other approach: the catalogue builds itself from the books, and lives on a server you open from any browser.

How Lexarii catalogues

Upload your EPUBs and PDFs, a few or a few hundred, and Lexarii:

• files each book by title, author, series and subject, read from the book itself, with its cover;
• groups one book's formats and editions into a single entry;
• gives you a subject list to browse what you own;
• keeps a file once, however many times it is uploaded;
• never invents a title. The publisher's details come first, and anything suggested to fill a gap is marked as a suggestion.

MOBI, AZW3, CBZ and CBR files are catalogued and shelved beside the rest, though the reader cannot open them yet. Your files are never altered.

Once catalogued, the library is also an OPDS catalogue, so KOReader, Thorium, Foliate and most other reading apps can browse it by author, series and subject too.

Try it

The demo has a small catalogued shelf of public-domain classics, reset every hour. To catalogue your own, subscribe or run it yourself, free.