.ODT File Format – Text content

Headings and paragraphs

The content of both is text mixed with the inline elements below, frames and shapes anchored to the paragraph, fields, index marks, bookmarks, notes, annotations (office:annotation, office:annotation-end) and change marks. The paragraph’s own text is its character data in document order, leaving out note bodies, ruby text, annotations and the content of shapes and frames.

Spaces, tabs and line breaks

Inside text:p and text:h, white space is collapsed. A reader must get the result of these steps, or a variant (below) that matches files from ODF 1.2 writers:

1. Replace each text:ruby with its text:ruby-base.
2. Remove descendants that are not text:s, text:tab or text:line-break and that cannot themselves contain them.
3. Replace descendants that can contain them (spans, links and so on) with their content.
4. Turn tab (U+0009), carriage return (U+000D) and line feed (U+000A) characters into spaces.
5. Remove spaces at the start and end.
6. Collapse each run of spaces into one.
7. Read the remaining text:s, text:tab and text:line-break as the characters they stand for.

The variant replaces step 2: first remove mark elements (change, bookmark, reference and index marks), then replace each other removable element with an imagined <text:s text:c="0"/>, so that spaces either side of it are not joined.

A soft hyphen is the character U+00AD, a non-breaking hyphen U+2011, and a non-breaking space U+00A0; ODF has no elements for them.

<text:p text:style-name="P1">Two<text:s/> spaces,<text:tab/>a tab and<text:line-break/>a new line.</text:p>

Spans, links and inline metadata

Lists

A list is text:list, with an optional text:list-header and then text:list-item elements, each holding paragraphs, headings or another list. The level of a list is how deeply it is nested: 1 at the top, and back to 1 inside a table cell or text box. A top-level list is one counting domain for its items and sub-lists. Whether numbers or bullets show depends on the list style.

• text:continue-list gives the xml:id of a list whose numbering this one continues. It overrides text:continue-numbering, which, when true, continues from the list just before if it has the same style.
• A list with no style takes its parent list’s style; failing that, a list style set by the paragraph styles inside it; failing that, an implementation’s default.
• If the list style has no style for an item’s level, the parent list’s style for that level is used, then the default list style.
• An item that only starts another list and has no text shows no number.

<text:list xml:id="list1" text:style-name="L1">
 <text:list-item><text:p>First</text:p></text:list-item>
 <text:list-item><text:p>Second</text:p>
  <text:list><text:list-item><text:p>Second, part a</text:p></text:list-item></text:list>
 </text:list-item>
</text:list>

Sections and page sequences

• text:display: true shown (default), none hidden, condition hidden when text:condition holds. A condition is a namespace prefix, a colon and a formula, such as ooow:Page > 1.
• text:protected: true stops the user editing the section (in the user interface only).
• text:protection-key: if present, a hash of the password (UTF-8) needed to remove the protection, made with the algorithm in text:protection-key-digest-algorithm: by default http://www.w3.org/2000/09/xmldsig#sha1; writers should use …#sha256.
• A linked section should still contain a full copy of the linked content.

Tracked changes

Changes are recorded in a text:tracked-changes element at the start of the scope they cover (usually office:text), and marked in the text by change marks that point at them with text:change-id. The specification warns that change tracking was under-specified in ODF 1.2 and that programs differ.

To restore deleted text at a text:change inside a paragraph, insert the deleted content without the opening tags up to its first text:p and the closing tags from its last </text:p> or </text:h>, so the first and last deleted paragraphs merge with the paragraph around the mark (if the last was a heading, the paragraph becomes one). Inside a heading, the same with text:h. Elsewhere, insert the deleted content as it is.

<text:tracked-changes>
 <text:changed-region xml:id="ct1" text:id="ct1">
  <text:deletion>
   <office:change-info><dc:creator>Ann</dc:creator><dc:date>2026-11-12T10:00:00</dc:date></office:change-info>
   <text:p>old </text:p>
  </text:deletion>
 </text:changed-region>
</text:tracked-changes>
<text:p>The <text:change text:change-id="ct1"/>text.</text:p>

Bookmarks and reference marks

Both come as a single point or as a start and end pair, matched by text:name, which must be unique. Ranges may overlap. Bookmarks are what hyperlinks and cross-reference fields point at; reference marks are what reference fields point at.

Footnotes and endnotes

A note sits in the text where it is anchored. How notes are numbered and where they are printed is set by text:notes-configuration in office:styles.

<text:p>A claim.<text:note text:id="ftn1" text:note-class="footnote">
 <text:note-citation>1</text:note-citation>
 <text:note-body><text:p text:style-name="Footnote">The source.</text:p></text:note-body>
</text:note></text:p>

Declarations

These come before the main content of office:text and define things the text uses: text:variable-decl (variable fields), text:user-field-decl (user fields), text:sequence-decl (sequence fields such as figure numbers), text:dde-connection (DDE fields and sections) and text:alphabetical-index-auto-mark-file (a file of words to index). Fields themselves are on the fields and indexes page.