詳細ガイド 近日公開
Markdown to Word Count Converterの包括的な教育ガイドを準備中です。ステップバイステップの解説、数式、実例、専門家のヒントをお届けしますので、もうしばらくお待ちください。
The Markdown to Word Count Converter strips Markdown syntax (headings `#`, emphasis `**bold**` and `*italic*`, links `[text](url)`, fenced code blocks ` ``` `, inline code `` ` ``, blockquotes `>`, lists `-`/`*`/`1.`, horizontal rules `---`, HTML tags) and returns accurate counts of words, characters (with and without spaces), sentences, paragraphs, plus estimated reading time at configurable speeds (default 225 WPM — Medium's industry-standard reading speed). Why strip Markdown first: raw character counts on Markdown source inflate by 10–30% from syntax (`**bold**` adds 4 characters, `[link](https://example.com)` adds 16+ characters that aren't part of the prose). Stripping yields counts that match what a reader actually sees on the rendered page — the metric writers, editors, and SEO tools care about. The converter uses a CommonMark-compliant tokenizer, so output matches what most renderers (GitHub, VS Code, Obsidian, Notion's export) produce. Reading time uses words ÷ WPM. Defaults: 225 WPM (Medium/standard publication, average adult reading prose at moderate pace), 200 WPM (web/SEO time-on-page estimate — Google's data suggests slightly slower reading on screens), 150 WPM (technical content with dense vocabulary), 300 WPM (light fiction or skimming). Sentence and paragraph counts help writers respect publication style guides that limit paragraph length (5–7 sentences for blogs, 2–3 for newsletters). Who this is for: Bloggers checking SEO-friendly post length (Google rewards 1,500–2,500 words for most topics), freelance writers billing per word, students respecting essay limits, content marketers planning newsletter segments, novelists tracking daily word goals, and technical writers checking documentation density. The tool runs entirely in-browser; paste sensitive drafts safely.
- 1Step 1 — Paste your Markdown into the input area
- 2Step 2 — Calculator strips syntax: `#` headings, `**bold**`/`*italic*`, `[links](url)`, fenced code blocks, inline code, blockquotes, lists, HTML tags
- 3Step 3 — Counts words by splitting stripped text on whitespace and filtering empty tokens
- 4Step 4 — Counts characters two ways: with spaces (total) and without spaces (typographic content)
- 5Step 5 — Counts sentences by detecting terminal punctuation outside abbreviations and decimals
- 6Step 6 — Counts paragraphs as blocks separated by blank lines after stripping
- 7Step 7 — Computes reading time = words ÷ configurable WPM (default 225)
Strips `#`, `**`, counts only prose tokens.
Typical SEO-friendly long-form post length.
Technical material slows readers — use lower WPM for accurate estimates.
Blog post SEO targets (1,500–2,500 words is typical sweet spot)
Freelance per-word billing verification
Academic essay word limits
Newsletter segment length planning
Documentation density audits
Daily writing goals (NaNoWriMo, daily journal)
What reading speed should I use for my content type?
Standard prose / blog posts: 225 WPM (Medium default). SEO time-on-page estimates: 200 WPM (Google data suggests slightly slower on screens). Technical / dense content: 150 WPM. Light fiction / skimming: 300 WPM. Audiobook narration: 150–160 WPM. Match WPM to your audience and content density.
Why does the converter strip Markdown before counting?
Raw Markdown source inflates character counts by 10–30% from syntax characters that readers don't see. `**bold**` adds four characters, `[link](https://example.com)` adds 16+. Stripping yields counts that match the rendered page — what writers, editors, and SEO tools actually care about.
How does sentence detection work?
The tokenizer looks for terminal punctuation (. ! ?) followed by whitespace and a capital letter. It uses heuristics to ignore common abbreviations (Mr., Dr., etc., e.g.) and decimal numbers (3.14). Sentence counts are approximate for ambiguous cases like quoted dialogue or run-on lists.
Does the tool count code blocks?
By default, fenced code blocks are stripped from the prose count because they're not natural-language reading. Inline code is also stripped. If you want code included, paste it outside fence markers.
Why might my count differ from Medium or Google Docs?
Each tool defines 'word' slightly differently. Medium counts hyphenated compounds as one word; Google Docs may count them as two. Markdown handling also varies. This calculator follows CommonMark spec and standard whitespace tokenization for consistency.
プロのヒント
Set reading speed to 200 WPM for SEO time-on-page estimates — Google's data suggests slightly slower reading on web than print. For technical writing or dense academic content, drop to 150 WPM. Communicating an honest reading-time estimate at the top of long posts measurably reduces bounce rate.