Skip to main content
Analysis

Markdown Structure Analyzer

Count Markdown words, UTF-16 characters, lines, blank-line paragraphs, ATX headings, links/images, fenced code blocks, and estimated reading minutes locally. Useful for writers tracking document length, editors reviewing content scope, and developers estimating reading time for documentation.

Markdown tools

In brief

This fast regex-based analyzer reports editorial indicators, not a full Markdown AST: code is excluded from words, images count with links, and characters are JavaScript UTF-16 units.

What this tool supports

  • Counts words after removing fenced/inline code, images, Markdown punctuation, and link destinations while retaining link labels.
  • Reports source characters, non-whitespace characters, physical lines, blank-line-separated paragraphs, and ATX headings using # through ###### followed by text.
  • Counts both inline links and images with one shared pattern and estimates fenced code blocks as pairs of triple-backtick markers.
  • Estimates reading time by ceiling(words / 220), with a minimum of one minute for any non-empty word count; supports live or manual analysis and local file import.

How to use Markdown Structure Analyzer

Follow this implementation-specific workflow for Markdown Structure and Reading-Time Analyzer; each step reflects the controls and limits in the current browser workspace.

  1. 1

    Step 1

    Provide the exact source

    Paste, type, or import the exact Markdown source whose editorial structure you want to measure.

  2. 2

    Step 2

    Choose the operation

    Keep Live enabled for continuous counts or disable it and use Analyze for a deliberate snapshot.

  3. 3

    Step 3

    Configure the available options

    Run the regex-based analyzer; do not assume it parses a complete Markdown abstract syntax tree.

  4. 4

    Step 4

    Run the local operation

    Review words, characters, lines, paragraphs, headings, links/images, code-fence pairs, and the 220-wpm estimate against the source.

  5. 5

    Step 5

    Verify and continue

    Use Markdown Preview to inspect rendered hierarchy and keep a recorded snapshot if statistics are part of an editorial gate.

Common mistakes

Treating the link count as links only

Consequence: The implemented pattern deliberately matches both [text](url) and ![alt](url), so images increase the same counter.

Fix: Inspect the source or a parsed preview when links and images need separate metrics.

Using character count as a Unicode grapheme count

Consequence: JavaScript string length counts UTF-16 code units; emoji and some combined characters can count as more than one visible symbol.

Fix: Use a grapheme-aware Unicode analyzer when user-perceived character limits matter.

Assuming paired backticks prove valid fenced code

Consequence: The code-block value is floor(number of ``` markers / 2), so malformed, nested, or language-edge fences can be miscounted.

Fix: Validate code fences in a real Markdown parser and use this number only as an editorial indicator.

Best practices

  • Compare statistics with a rendered preview before enforcing publication rules.
  • Document the 220 words/minute assumption if the reading estimate is shown to users.
  • Use stable source text and parser-aware tooling for contractual limits; this analyzer is intentionally lightweight.

Connected workflows

Editorial Markdown quality check

Markdown Structure AnalyzerMarkdown Live PreviewMarkdown to HTML Viewer

Measure a stable source snapshot, inspect the rendered hierarchy, then convert the approved Markdown into sanitized HTML.

  1. Freeze the source revision
  2. Analyze editorial indicators
  3. Inspect rendered headings and links
  4. Correct source issues
  5. Generate sanitized HTML

Browser limitations

  • Word and structure detection uses regular expressions rather than a Markdown parser, so escaped syntax, HTML blocks, reference links, indented code, and unusual fences can be misclassified.
  • Characters are UTF-16 units, paragraphs are blank-line groups, and only ATX # headings are counted; Setext headings are omitted.
  • The entire source is processed synchronously on the main thread and no report download or historical comparison is provided.

Authoritative resources

CommonMark Specification 0.31.2Publisher: CommonMark · Reviewed: 2026-07-26

Concrete examples

Input
# Title Hello world. [Docs](https://example.com)
Result
The analyzer counts one ATX heading, one inline link, three blank-line groups, and four plain-text words.

The link destination is removed while “Docs” remains a word; the non-empty four-word total rounds up to one minute at 220 words per minute.

Frequently asked questions

Does the word count include fenced and inline code?

No. Triple-backtick blocks and single-backtick spans are replaced with spaces before words are split.

Why is an image included in the Links card?

The current regular expression uses an optional exclamation mark, so both image and link inline syntax share one count.

Are underlined Setext headings counted?

No. headingOutline selects only lines matching one to six # characters followed by whitespace and non-whitespace text.

What does “characters” mean for emoji?

It is JavaScript string length in UTF-16 code units, not Unicode code points or user-perceived grapheme clusters.

How is reading time calculated?

Zero words gives zero; otherwise the tool computes ceil(words / 220) and never reports less than one minute.

Privacy and offline behavior

Markdown is analyzed only in this browser and is neither uploaded nor persisted by this workspace. Imported files are read after selection, and analysis works offline once the application bundle is cached.