What is an FB2 file? FictionBook, explained

A file ending in .fb2 is a FictionBook ebook — and unlike almost every other ebook format, it is not an archive. There is no zip to open, no folder of parts, no manifest. An FB2 file is one XML document: the metadata, the chapters, and even the images, base64-encoded, all sitting in a single text file you can open in an editor and read.

If an FB2 file has just landed on you — from an old backup or a friend’s collection — the practical questions are what it is, what still opens it, and what happens to the book when you convert it to something more widely supported. We can answer the last one precisely, because we ran the conversion on a real file while writing this and kept the numbers.

One XML document, no archive

Here is the actual shape of an FB2 file — this excerpt is from the specimen we built for this post from our own test text, trimmed but otherwise verbatim:

<?xml version="1.0" encoding="utf-8"?>
<FictionBook xmlns="http://www.gribuser.ru/xml/fictionbook/2.0"
             xmlns:l="http://www.w3.org/1999/xlink">
  <description>
    <title-info>
      <author>
        <first-name>Specimen</first-name>
        <last-name>Author</last-name>
      </author>
      <book-title>The Conversion Specimen</book-title>
      <lang>en</lang>
    </title-info>
  </description>
  <body>
    <section>
      <title><p>CHAPTER 1: WHAT THIS FILE IS</p></title>
      <p>This is the specimen book Epub Studio uses…</p>
    </section>
  </body>
  <binary id="cover.png" content-type="image/png">iVBORw0KGgo…</binary>
</FictionBook>

Three things in that structure explain most of what FB2 does well and badly.

First, the tagging is semantic. A <section> is a chapter, a <title> is its heading, a <p> is a paragraph. There is no font, no page size, no layout instruction anywhere — the reading app decides all of that. In this respect FB2 and EPUB agree completely; both are reflowable formats, and which format a book should be in is a separate question from which container it sits in.

Second, the metadata lives inside the same file, in the <description> block, in named fields — not in filename conventions, not in a sidecar. Any tool that speaks XML can read the author and title without guessing.

Third, the images are text. A <binary> element holds each image as base64, which inflates it by a third: the 56,542-byte PNG we embedded in the specimen became 75,392 characters of encoded text. That is the price of the single-file design, and it is why an FB2 with pictures is noticeably larger than the same book packed as EPUB — our specimen was 79,032 bytes as FB2 and 57,561 bytes after conversion, with identical content.

Compare that with EPUB, which is a zip archive of separate XHTML, CSS and image files with a manifest — we have opened one up piece by piece before. The archive design costs complexity but stores images as images. The FB2 design costs bytes but means the book is never “corrupt because the zip broke”.

How to open an FB2 file today

Honestly: it depends on the reader you already use, and we are not going to publish a support table we have not tested. Some reading apps open .fb2 directly; many list only EPUB, PDF and a couple of others in their import dialogs. The one route we can vouch for first-hand is the file itself — FB2 is plain text, so a text editor or a browser will show you the book’s content on any platform, unstyled but readable, which is more than most formats can say.

If the file arrives zipped — .fb2.zip is common precisely because XML compresses so well — unzip it first. The zip is a wrapper, not part of the format, and tools that accept .fb2 (ours included) want the plain file from inside it.

For actually reading at length, conversion to EPUB is the pragmatic answer, because EPUB is what nearly everything opens. Which brings us to what conversion does to the book.

We converted one — here is what survived

We built a real FB2 from our specimen text — four chapters, an emphasis tag inside one paragraph, a footnote body, and an embedded 56,542-byte PNG — and dropped it on our FB2 to EPUB converter. The conversion runs in the browser; the file never leaves your device.

What the output actually contained, checked entry by entry:

The losses are the honest part of that table, so to spell them out: the converter reads each paragraph’s text and rebuilds it clean, which means inline markup inside paragraphs — emphasis, strong, inline links — is flattened to plain words. The separate notes body, FB2’s home for footnotes, is skipped entirely. The current build also repeats each section title as the chapter’s first paragraph — a cosmetic duplication we found while testing for this post and have logged. And although the embedded image is decoded back to an exact copy of the original PNG inside the EPUB, nothing marks it as the cover, so a shelf app will show a blank tile — the same symptom, from a different cause, as a cover that exists but is not declared, and the same one-minute fix in the cover editor.

For a plain prose novel, none of those losses touches the reading experience. For a footnoted or heavily formatted book, they do, and a full desktop converter such as Calibre is the right tool for that job. That is not a hedge; it is what the test showed.

The result is structurally sound, though. The converted specimen went straight through our EPUB validator:

One more first-hand data point: a malformed FB2 fails fast. We deleted a single closing tag from the specimen and the converter refused it with “This FB2 file has malformed XML and cannot be read.” — XML’s strictness is the format’s safety rail here, exactly as it is inside an EPUB’s own chapter files.

Keep the FB2 or convert it?

Keep the original either way — it is small, it is text, and it archives well. But for day-to-day reading the arithmetic favours EPUB: every mainstream reading system opens it, it stores images without the base64 tax, and a converted file that validates cleanly will behave in strict readers too. Drop your file on the FB2 to EPUB converter, give the output a cover if the book had one, and it is a normal ebook from then on. If what you actually need is a PDF for printing, convert to EPUB first and run the result through EPUB to PDF — two steps, because each one is doing a genuinely different job.