MD044 - Use Proper Capitalization for Names¶

Aliases: proper-names

What this rule does¶

Ensures brand names, product names, and technical terms are consistently capitalized throughout your documents.

Why this matters¶

Professional appearance: Incorrect capitalization looks careless and unprofessional
Brand respect: Shows you care about getting company and product names right
Consistency: Readers won't be confused by different variations of the same name
Searchability: Consistent naming improves search results and indexing

Examples¶

✅ Correct¶

With configured names:

[MD044]
names = ["JavaScript", "GitHub", "Node.js", "TypeScript"]

# JavaScript Development Guide

This guide covers JavaScript best practices for GitHub projects.

We'll be using Node.js and TypeScript for our examples.

❌ Incorrect¶

# Javascript Development Guide        <!-- Should be "JavaScript" -->

This guide covers javascript best practices for github projects.
<!-- "javascript" and "github" are incorrectly capitalized -->

We'll be using NodeJS and typescript for our examples.
<!-- "NodeJS" should be "Node.js", "typescript" should be "TypeScript" -->

🔧 Fixed¶

# JavaScript Development Guide

This guide covers JavaScript best practices for GitHub projects.

We'll be using Node.js and TypeScript for our examples.

Configuration¶

[MD044]
names = []                       # List of properly capitalized names
code-blocks = false              # Check inside code blocks (default: false)
html-elements = true             # Check inside HTML elements (default: true)
html-comments = true             # Check inside HTML comments (default: true)
whole-words = false              # Preserve names inside compound identifiers (default: false)
ignore-frontmatter-fields = []   # Frontmatter fields to skip (default: check all)

Automatic fixes¶

When enabled, this rule will:

Replace all incorrect capitalizations with the proper version
Preserve the context and meaning of your text
Work across headings, paragraphs, and lists

Special cases¶

Only checks names you've configured
Set whole-words = true to match a configured name only when it occupies a whole prose token. With this option, runs of hyphens and underscores between Unicode letters or numbers join identifier components, so abc is checked in (abc) and [abc] but not in 123-abc-def, foo_abc_bar, foo__abc__bar, or Unicode compounds such as café_abc_menu. Underscores parsed as Markdown emphasis delimiters do not join tokens: _abc_, __abc__, and the emphasized abc in foo-__abc__-bar are still checked. Emphasizing an entire identifier, such as _foo_abc_bar_ or _abc_def_, still protects its embedded components. The default is false for markdownlint compatibility, so compound components continue to be checked unless this option is enabled. This setting does not add a general path or URL exemption; existing link and URL exclusions still apply.
Won't check inside URLs
Preserves filename extension suffixes such as favorite.toml, .toml, and config.toml.example when TOML is configured. This applies in both body text and frontmatter without an additional setting. Only the suffix is exempt: standalone node.js is still checked against the complete name Node.js, and a configured name in a filename stem such as myapp.png is still checked. The dot is a syntax cue, so dotfiles such as .github and dotted suffixes such as api.node.js also retain their spelling. A missing space after a sentence period can therefore hide a capitalization error; text after ellipses or a period followed by a space is still checked
Won't check inside a file path in a frontmatter value, such as docs/myapp.md, ./myapp.md, or /abs/path/myapp.md, since rewriting a name there would break the reference. A slash is required for this exemption: a slash-less value like image: myapp.png has no path signal, so a configured name in its stem is still flagged and fixed. This whole-path exemption applies to frontmatter values only; names in body-prose paths are still checked, apart from extension suffixes. A path is recognized when it has a leading ./, ../, ~/, or /, or its final segment has a file extension. A bare relative path with a literal space and no extension (docs/My App/myapp) is not recognized and is still checked
Won't check inside a wiki embed (![[myapp.png]], ![[myapp.png|300]]). The whole construct names a file, so rewriting any part of it changes what the embed resolves to
In a wikilink, checks what the reader sees: the display text of [[page|the myapp guide]], and the page name of a plain [[myapp]], which doubles as its own display text. The page name in front of a | is left alone, since renaming it changes what the link resolves to without changing anything on the page
Won't check inside a template shortcode tag ({{% nodejs %}}, {{< figure alt="nodejs" >}}). A shortcode is resolved by name against a template, so the text in the tag names what to invoke rather than mentioning a product, and rewriting it points the invocation at a template that does not exist. Only the tag itself is exempt: prose on the same line is still checked, and so is the body between a paired {{% note %}} and {{% /note %}}, which the site generator renders as ordinary Markdown
By default, skips code blocks and inline code. Wrap function calls such as oklch(.5 .5 255) in backticks to preserve their spelling. Unformatted calls are checked like prose. Set code-blocks = true to check both code blocks and inline code
Set html-comments = false to skip HTML comment content
Some names are intentionally lowercase (like "npm")

Frontmatter¶

rumdl checks proper names inside frontmatter values, not just body text, since fields like title and description often hold prose. Frontmatter keys themselves are never checked.

To exclude specific fields (for example machine-readable slugs or SEO metadata that happen to contain a configured name), list their top-level keys in ignore-frontmatter-fields:

[MD044]
names = ["MyApp"]
ignore-frontmatter-fields = ["slug", "seo"]

Field names are matched case-insensitively against top-level frontmatter keys only. Excluding a parent key excludes everything nested under it, including nested maps, sequence items, and multi-line values. In TOML frontmatter, excluding a key also excludes its [table] and [[array-of-tables]] sections.

One YAML shape is not covered: a flow collection whose contents continue on a following line that begins at column zero (seo: [ then myapp on the next line) attributes that line to its own text rather than to seo, so ignore-frontmatter-fields = ["seo"] does not exclude it. The value keeps being checked, which is the safe direction. Put the collection on a single line, or indent its continuation, to have it excluded.

Learn more¶

MD033 - Control HTML usage in Markdown
MD040 - Identify code block languages
MD049 - Use consistent emphasis markers