Free Tool · Baidu SEO · Chinese Text Analysis

Chinese Word Segmentation Tool (分词)

Chinese has no spaces between words. Paste a text to see how it splits into words — with Pinyin, word frequency, and a check of how your SEO keyword is split, the way search engines like Baidu match it.

Split Chinese Text into Words

  • Try:

Up to 10,000 characters. Results update as you type.

Keyword check

外贸独立站 → 外贸 / 独立站

Split into 2 words. The full keyword appears 1× in your text · 外贸 1× as a word · 独立站 1× as a word.

66 words · 59 different words · 128 Chinese characters · average 1.9 characters per Chinese word

Segmented text
wàimào外贸dúlìzhàn独立站xiǎngzài想在bǎidù百度hé和gǔgē谷歌huòdé获得zìránliúliàng自然流量，guānjiàncí关键词yánjiū研究shì是dì第yībù一步。hěnduō很多qǐyè企业zhǐshì只是bǎ把yīngwén英文guānjiàncí关键词fānyì翻译chéng成zhōngwén中文，què却hūlüè忽略le了zhōngguó中国yònghù用户zhēnshí真实de的sōusuǒxíguàn搜索习惯。wǒmen我们jiànyì建议xiānyòng先用bǎidùzhǐshù百度指数fēnxī分析sōusuǒliàng搜索量，zài再jiéhé结合xiǎohóngshū小红书hé和dǒuyīn抖音shàng上de的rèmén热门huàtí话题，xiě写chū出fúhé符合yònghùxūqiú用户需求de的zhōngwén中文nèiróng内容。nánjīngshì南京市chángjiāng长江dàqiáo大桥de的lìzi例子shuōmíng说明，tóng同yījùhuà一句话kěyǐ可以yǒu有bùtóng不同de的fēncí分词fāngshì方式。

Most frequent words (without function words such as 的, 了, 和)

WordPinyinCountShare
关键词guānjiàncí23.0%
中文zhōngwén23.0%
外贸wàimào11.5%
独立站dúlìzhàn11.5%
想在xiǎngzài11.5%
百度bǎidù11.5%
谷歌gǔgē11.5%
获得huòdé11.5%
自然流量zìránliúliàng11.5%
研究yánjiū11.5%
第dì11.5%
一步yībù11.5%
很多hěnduō11.5%
企业qǐyè11.5%
只是zhǐshì11.5%
英文yīngwén11.5%
翻译fānyì11.5%
成chéng11.5%
却què11.5%
忽略hūlüè11.5%

Next Tools for Chinese SEO

Chinese SEO Content Editor Analyse the top Baidu pages for a keyword and score your draft as you write. Baidu SERP Snippet Preview See how your title and description appear on Baidu, with keywords highlighted. Pinyin URL Slug Generator Word-based Pinyin URLs — the same segmentation, turned into clean slugs.

All free China SEO and web tools →

What Is Chinese Word Segmentation?

English separates words with spaces; Chinese doesn't. 我们提供中文搜索引擎优化服务 is a continuous string of characters, and a reader — or a search engine — has to decide where each word ends: 我们 / 提供 / 中文 / 搜索 / 引擎 / 优化 / 服务. This step is called word segmentation, or 中文分词.

Most of the time the split is obvious to a native speaker but not to software. The classic example is 南京市长江大桥: correctly 南京市 / 长江大桥, "the Yangtze River Bridge of Nanjing City" — but a naive split produces 南京 / 市长 / 江大桥, "Nanjing mayor Jiang Daqiao". Every Chinese search engine, translation system and text analysis tool starts with this step.

Why Segmentation Matters for Baidu SEO

  • Baidu matches words, not strings. A query is split into words, and pages are matched on those words. That is why a page can rank for 外贸独立站推广 without containing the exact phrase — and why Baidu highlights separate parts of a keyword in red.
  • Long-tail keywords are combinations. 跨境电商 / 独立站 / 推广 / 方法 — seeing the parts helps you cover a topic with the words people combine in their searches.
  • Translations miss words. An English keyword translated literally often produces a word combination Chinese users don't search for. The word frequency list shows which terms a text actually focuses on.
  • Brand names get split. A new brand name made of common characters may be split into ordinary words. Using it consistently — and in Latin letters where it helps — makes it recognisable.

How This Tool Segments Text

The tool runs on our server. It uses the ICU word-boundary dictionary, which is built into browsers and operating systems, and refines it with a list of about 47,000 common Chinese words plus current marketing terms such as 小红书, 抖音, 独立站 and 种草 — words a general dictionary often splits apart. Pinyin comes from the same data as our Chinese to Pinyin converter.

Baidu's segmenter is proprietary and uses much larger models, so treat the result as a close approximation: everyday text splits the same way; brand names, new words and slang may differ. Developers building a site search for Chinese content will find the same principle in open-source segmenters such as jieba (Python) or IK Analyzer for Elasticsearch.

Before you publish

Chinese Keywords and Segmentation: Checklist

Use these points when you optimise a Chinese page for a keyword — whether you write it yourself or check a translation.

  1. Put the full keyword in the title

    Baidu can match the parts, but the complete phrase in the title — near the start — is still the strongest signal. Preview the title on Baidu →

  2. Use every part of the keyword naturally in the text

    Check the keyword's split in this tool, then make sure each part appears in headings and body text, in context — not as a repeated block.

  3. Don't stuff

    If one word takes a large share of the frequency list, the text reads like keyword stuffing. Use related words and synonyms instead.

  4. Never add spaces between Chinese words

    Some translated texts insert spaces to "help" search engines. They don't help, look unnatural, and can change how text is matched.

  5. Research how Chinese users phrase the query

    Translate the intent, not the English keyword. Baidu Index and Baidu's keyword planner show which word combinations people actually search for. Our team's toolkit →

  6. Use a Chinese-aware site search

    A site search that splits on spaces finds almost nothing in Chinese. Use a search engine configuration with a Chinese segmenter.

From our Baidu SEO and Chinese content work for Western companies since 2012.

Get a Chinese keyword strategy

Frequently Asked Questions

Chinese is written without spaces between words, so software has to work out where one word ends and the next begins. This step — 中文分词 — is the basis of Chinese search engines, machine translation, text analysis and site search. 百度推广 is two words, 百度 and 推广; 研究生命起源 is 研究 / 生命 / 起源, not 研究生 / 命 / 起源.
Baidu splits both search queries and web pages into words and matches them, which is why results can rank for a query that never appears as one exact phrase, and why Baidu highlights individual parts of a keyword in red. Baidu's own segmenter is proprietary, but it works on the same principle as this tool.
Not always. This tool combines the ICU word-boundary dictionary built into many systems with a list of about 47,000 common Chinese words and current marketing terms. Baidu uses its own, much larger models. For most everyday text the results are close; brand names, new words and slang may be split differently.
Because Baidu matches words, not character strings. A keyword such as 外贸独立站 is split into 外贸 and 独立站; a page that uses the full phrase in the title and both parts naturally in the text matches it well. Segmentation also shows which words dominate a text — a quick check of keyword focus.
南京市长江大桥 means "the Yangtze River Bridge of Nanjing City" (南京市 / 长江大桥). A naive split reads it as 南京 / 市长 / 江大桥 — "Nanjing mayor Jiang Daqiao". It shows why segmentation needs a good dictionary and context.
Yes, mostly. The ICU dictionary covers Traditional characters, and the Pinyin data includes them. The added word list is Simplified, so Traditional text may split into slightly smaller pieces.
Yes, it is free and needs no sign-up. Your text is only used to show the result and is not stored or logged. You can paste up to 10,000 characters at a time.

Built by Jademond Digital, a Chinese digital marketing agency founded in Guangzhou in 2012. Word list: mozillazg/phrase-pinyin-data (MIT licence). About us →