Standard Ebooks Tools

Last reviewed Sep 13, 2026 · 776 words

Nothing in the Standard Ebooks tools serves a web page. The repository is the se command-line toolchain the project uses to turn public-domain texts into the best-typeset EPUBs available anywhere, and the "self-hosted ebook library" the catalogue alludes to is a different program, Calibre-Web or Kavita, that you would point at the files se produces. Read this page as "how a self-hoster uses a production tool", because that is what it is: 1,666 stars' worth of typography obsession, and the only reason to install it is that you intend to make or fix ebooks rather than just read them.

What the se command actually does

se create-draft scaffolds a new ebook project with the Standard Ebooks directory layout, metadata skeleton and CSS. se typogrify rewrites straight quotes, dashes, ellipses and spacing to proper typographic characters across every XHTML file. se lint runs several hundred checks, from metadata consistency to semantic markup to whether a section break is coded correctly, and prints a numbered list of everything the Standard Ebooks manual of style would reject. se build compiles the source into a compatible EPUB, an advanced EPUB 3, and optionally Kindle and Kobo variants, with the covers and metadata generated along the way. There are dozens more subcommands for word counts, spelling modernisation, semantic tagging and checking links, all documented at standardebooks.org/tools. The command line is the interface; there is no GUI and the project is not interested in adding one.

Requirements are heavier than the code suggests

Installation is pipx install standardebooks, which is the easy part. se build calls out to epubcheck, which needs a Java runtime, and building Kindle output needs Calibre's ebook-convert on the path, so a full setup means Java, Calibre and a few image tools alongside. The catalogue's 256 MB figure is about the tool itself; add Java and a Calibre install and you want a real workstation or a 2 GB container, not a Raspberry Pi Zero. The tools target Linux and macOS and are actively developed against current Python; on Windows, use WSL rather than fighting it.

It is for producing books, not converting your library

This is where expectations go wrong. se lint on an EPUB bought from a retailer will produce hundreds of errors, because the tool assumes a book authored to the Standard Ebooks style from the ground up, with semantic inflection, hand-checked metadata and a specific file layout. It is not a batch fixer for a Calibre library, and Calibre itself remains the tool for conversions, metadata scraping and DRM-free format shuffling. Where the SE tools shine is producing a new edition of a public-domain text from a Project Gutenberg transcription: se create-draft, paste in the text, spend a few evenings on markup and proofreading, se typogrify, iterate on se lint until it is clean, se build. The result is an ebook that renders beautifully on every reader you own.

Pair it with a library server to actually read the output

The workflow that makes sense on a home server is a build machine plus a library. Keep your SE projects in Git, as the project does, build on your workstation, and drop the finished EPUBs into the directory your library server watches. Kavita and Komga handle EPUBs with good readers in the browser and OPDS feeds for phones. Calibre-Web adds Calibre's metadata database and Kindle sending. Any of the ebooks category options works; the SE tools do not care what serves the file.

Most readers should just download the catalogue

If your goal is a great-looking library rather than the craft of producing one, the Standard Ebooks website offers every finished title as a free download, and an OPDS feed and bulk downloads exist as a patron perk at last check. A few hundred exquisitely typeset classics in your Kavita library is 20 minutes of work that way, and zero installs of Java.

What I'd do

I would install the tools on my desktop, not my server, and use them for exactly one thing: producing my own editions of public-domain texts I care about, following the manual of style, and contributing the good ones back upstream. The output would go into Kavita on the home server, next to the Standard Ebooks catalogue I downloaded the easy way. I would not try to run se lint over a purchased library, and I would not expect the tools to replace Calibre for anything that is not a Standard-Ebooks-style project.