Chinese Sensitive Words Mcp

What is Chinese Sensitive Words Mcp?

An MCP server that detects Chinese sensitive words and banned terms in text, supporting platform-specific word lists for Xiaohongshu, Douyin, Kuaishou, and Bilibili. It is designed for content creators, editors, and moderators who need to check compliance of copy, product descriptions, or livestream scripts.

How to use Chinese Sensitive Words Mcp?

Install via npx without registration (100 free checks per day) or with a token for unlimited use. Add a configuration entry to Claude Desktop, Cursor, Windsurf, or Claude Code. Invoke the check_sensitive_words tool to analyze text, or get_word_suggestions to get safe replacement suggestions for a given keyword.

Key features of Chinese Sensitive Words Mcp

• Extensive word list updated daily, covering politics, pornography, violence, gambling, drugs, advertising law, and medical claims.
• Multi-platform word libraries for Xiaohongshu, Douyin, Kuaishou, and Bilibili.
• Risk level classification: high (ban risk), medium (throttling risk), low (suggested modification), and tip.
• Replacement suggestions for flagged words.
• Homophone and variant character detection (e.g., "薇信" for "微信").
• Gap character detection (e.g., "加 微 信").
• Phone number and URL detection.
• NER-based intelligent filtering to reduce false positives on names and places.

Use cases of Chinese Sensitive Words Mcp

• Check social media posts for banned terms before publishing on Xiaohongshu or Douyin.
• Analyze livestream scripts for words that could cause throttling or account bans.
• Get safe alternative phrasing for high-risk words flagged during content review.
• Verify product descriptions comply with advertising law and medical efficacy regulations.
• Automate compliance checks in content moderation workflows.

FAQ from Chinese Sensitive Words Mcp

How is Chinese Sensitive Words Mcp different from other sensitive word checkers?

It offers multi-platform word libraries (Xiaohongshu, Douyin, Kuaishou, Bilibili), risk level classification, replacement suggestions, homophone detection, and free usage up to 100 times per day without registration.

What are the runtime dependencies?

It runs via npx and requires Node.js. No additional database or server setup is needed for the free tier.

Where does the data live and is it processed locally?

The server sends text to an external API for analysis. No data is stored by the MCP server itself.

Are there any usage limits?

The free tier allows 100 checks per day without a token. Registered users with a token have unlimited usage.

What transport and authentication does the server use?

The server uses stdio transport. Authentication is optional via an environment variable WORDSCHECK_ACCESS_TOKEN for unlimited access. You can also configure a custom API base URL with WORDSCHECK_API_BASE.