Each TOC entry carries a 4-byte ASCII type tag. This page documents every tag and describes the data each chunk contains.

CHAP Chapter

Tag: CHAP

The book’s chapter content. The data is either Markdown or HTML text. The markup_type field in the TOC entry determines which format.

TOC field Description
markup_type 0 tag “1” color:#3b82f6 HTML
compression Typically 1
language BCP 47 index for chapter language 
Chapter data: UTF-8 encoded Markdown or HTML string

IMG_ Image

Tag: IMG_

The trailing underscore distinguishes this tag from IMG-related variant tags.

An inline image referenced from chapter content.

TOC field Description
compression Typically 0
content_type Image format hint for JPEG, PNG, or WebP 
Image data: raw image bytes in JPEG, PNG, or WebP format

The content type is validated by magic bytes on read. Supported formats include JPEG, PNG, and WebP.

CSS_ Stylesheet

Tag: CSS_

A CSS stylesheet for rendering chapter content.

TOC field Description
compression Typically 1 
CSS data: UTF-8 encoded CSS string

The CSS is parsed and validated with cssparser on write.

FONT Font

Tag: FONT

An embedded font file.

TOC field Description
font_embedding Embedding mode such as subset or full
font_license_url_len Length of embedded license URL
compression Typically 0 
Font data: raw font bytes in WOFF2, TTF, or OTF format

The font format is detected by magic bytes:

Magic bytes Format
wOF2 or \x00\x01\x00\x00 WOFF2
OTTO OTF with CFF outlines
\x00\x01\x00\x00 TTF
true TTF in Mac format

The font license URL is a UTF-8 string stored as a variable-length field following the standard TOC fields.

COVR Cover Image

Tag: COVR

The full-size book cover image.

TOC field Description
cover_type 1 tag “2” color:#3b82f6 back cover
compression Typically 0 
Cover data: raw image bytes in JPEG, PNG, or WebP format

COVT Cover Thumbnail

Tag: COVT

An optional downsized version of COVR for quick previews in listing views or ebook reader shelves.

TOC field Description
cover_type 1 tag “2” color:#3b82f6 back cover
compression Typically 0 
Thumbnail data: raw image bytes in JPEG, PNG, or WebP format

NOTE Note or Annotation

Tag: NOTE

Embedded annotations associated with the file.

Note data: application-defined format

This chunk type stores per-file annotation data. For a richer annotation model that travels with the file, see the org.nisoku.anno EXTRA namespace.

SIDX Search Index

Tag: SIDX

An inverted search index for the book’s content.

SIDX MessagePack structure 
{
  "version": 1,
  "terms": {
    "word": [
      {"chunk": 0, "positions": [42, 105]},
      {"chunk": 2, "positions": [17]}
    ],
    "another": [
      {"chunk": 1, "positions": [88]}
    ]
  }
}

Terms are lowercase whitespace-split tokens. chunk refers to the TOC index of the chapter. positions are byte offsets within the decompressed chapter.

See the Search Index feature for usage details.

MATH Math Equation

Tag: MATH

A math equation embedded in the book.

TOC field Description
content_type Math format: 0 tag “1” AsciiMath 
Math data: UTF-8 encoded equation string

Equations are rendered at read time by the consuming application, for example with KaTeX or MathJax.

Next Steps

Wire Format
Compression
Layout Modes
Previous Wire Format Next Compression