字数统计与写作时长估算

Characters vs Words

These two terms point to different things across Chinese and English:

• Characters (字数): each character (Chinese, Latin letter, digit, punctuation) counts as one. In Chinese, 1 汉字 = 1 character.
• Words (词数): a run of letters or digits counts as one. In English, 1 word = 1 word; in Chinese, if segmented, "今天天气真好" is 4 words (今天/天气/真/好).

How mainstream tools handle this:

Our Word Counter shows character count, Chinese characters, English words, sentences, and paragraphs side-by-side, so mixed text isn't misread.

Counting Rules for Chinese & English

How Chinese characters are counted

Standard convention: the CJK Unified Ideographs range (U+4E00–U+9FFF) plus Extension A (U+3400–U+4DBF) are counted as Chinese characters. Chinese punctuation (，。！？；：""''…) and English punctuation are not counted as characters, but they do count toward total characters.

Common edge cases:

• Full-width vs half-width: full-width letters (ＡＢＣ) are treated as Chinese; half-width (ABC) as English. Inconsistent counting results if you mix them — use half-width uniformly.
• Traditional Chinese: traditional characters (漢字) are in the CJK range, counted normally.
• Punctuation: Chinese punctuation is excluded from character count, so a 100-character article may actually be 110 characters total.

How English words are counted

Mainstream rule: a run of letters or digits counts as one word. Supports abbreviations (Mary's = 1 word), numbers (3.14 = 1 word, 1,000 = 1 word), hyphenated compounds (well-known = 1 or 2 depending on the tool). Chinese punctuation does not affect English word detection — "Hello，world" is still 2 words.

Mixed text

"今天发布新 product！" under the mainstream rule: 5 Chinese characters (今天发布新) + 1 English word (product) + 1 exclamation. Report both numbers separately, then combine via reading time.

Reading & Speaking Speeds

"How long will it take to read?" is characters divided by speed. Baselines from research:

These are baselines; actual speed varies: complex content (academic, legal) is 30-50% slower; simple content (children's reading, conversational) can be 30% faster. Tool outputs are estimates, for reference not for evaluation.

Genre-Based Lengths

Same "article" can mean wildly different lengths by genre:

Decide a target length before writing, then check the gap afterward. A large gap is usually a structure problem — not more words or fewer, but re-allocated section weight.

Common Writing Targets

Social platforms

• WeChat Official Account: 800-2000 characters for body; title 10-30 characters; summary 40-80 characters.
• Zhihu answer: 500-2000 characters for dry-goods; 5000 characters for deep analysis.
• Xiaohongshu: 300-800 characters for body; title 10-20 characters.
• Twitter/X: 140 characters (~28-35 English words); long content uses threads.

Writing and academia

• Paper abstract: 200-300 words (English); ~400-600 Chinese characters.
• Thesis body: undergraduate 5000-15000 chars; master's 20000-50000 chars; PhD 50000+ chars.
• Email body: 30-200 characters; long emails should be bulleted.

Speeches

An 8-15 minute talk corresponds to 1500-3000 Chinese characters, or 1000-2000 English words. Leave a 10-15% buffer (live events tend to overrun).

Word Count in SEO

Google has explicitly stated: there is no hard word-count requirement. But actual data supports a correlation between content depth and rankings:

• Blog posts: top-ranking pages median about 1000-1500 English words (~2000-3000 Chinese characters).
• Programmatic pages (like our tool pages): short text is fine, provided there's structured data (JSON-LD) and semantically clear titles.
• Guides/tutorials: being close to "the actual problem the reader wants to solve" matters more; word count is a signal, not a goal.

Common mistakes:

• Padding to hit a target: dilutes information density and actually lowers rankings. SEO cares about the "information per character" ratio, not raw character count.
• Reading English word count for Chinese content: severely underestimates depth.
• Skipping structured data: word count is fine but without JSON-LD, Google Rich Results can't recognize it — missed rich-snippet opportunity.

Right approach: start from the problem the reader wants solved, then write at the genre's typical length; afterward, verify the word count actually covers the problem — don't reverse-engineer content from a word target.

FAQ