sysprog21

sysprog21 / zhtw-mcp

Public

A linguistic linter for Traditional Chinese (zh-TW)

claude-code large-language-models linguistic-analysis localization-tool traditional-chinese
91
6
100% credibility
Found Mar 10, 2026 at 60 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Rust
AI Summary

A tool that checks and corrects AI-generated Traditional Chinese text to follow official Taiwan standards for words, punctuation, and characters, integrating seamlessly with AI coding assistants.

How It Works

1
👀 Notice odd Chinese words

While using AI to write in Traditional Chinese, you spot unfamiliar words or punctuation that don't feel right for Taiwan style.

2
📥 Get the helper tool

Download this free checker that spots and fixes Taiwan Chinese issues in your AI text.

3
🔗 Connect to your AI chat

With one simple setup, link it to your AI assistant like Claude so it works automatically inside your chats.

4
🔍 Check your writing

Paste any text or tell your AI 'Check this for Taiwan Chinese style' and it scans instantly.

5
✨ See smart fixes

Get a clear list of issues with easy suggestions, like changing mainland words to Taiwan ones or fixing punctuation.

6
✅ Apply and perfect

Click to fix everything automatically, or use it in your files for clean, proper Taiwan Chinese every time.

🎉 Write confidently

Your AI now produces authentic Taiwan-style Chinese, saving time and avoiding mix-ups.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 60 to 91 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is zhtw-mcp?

zhtw-mcp is a Rust-built linguistic linter for Traditional Chinese (zh-TW) that enforces Taiwan Ministry of Education standards on vocabulary, punctuation, and character shapes. It scans text for Mainland Chinese (zh-CN) influences—like "軟件" instead of "軟體"—and auto-fixes them via CLI commands like `zhtw-mcp lint --fix file.md`. As an MCP server, it plugs directly into AI coding assistants like Claude Code, catching regional drift in generated zh-TW docs or comments without extra API keys.

Why is it gaining traction?

AI models trained on web data overwhelmingly favor zh-CN terms, even for zh-TW prompts, making clean Traditional Chinese output unreliable. This linter stands out by compiling 1000+ rules into a fast binary, offering profiles like strict_moe for punctuation and ui_strings for casing (JavaScript, GitHub), plus Markdown-aware scanning that skips code blocks. Developers hook it via simple MCP registration or CLI, getting instant fixes that converge in one pass.

Who should use this?

Taiwan software engineers writing READMEs, docs, or UI strings in zh-TW, especially those relying on AI for content generation. Content teams localizing tech blogs or open-source repos to Traditional Chinese standards. DevOps folks integrating it into CI/CD for zh-TW quality gates on pull requests.

Verdict

Grab it if you write zh-TW—solid docs, CLI, and MCP integration make it dead simple despite 15 stars and 1.0% credibility score. Early-stage with room for more rules, but MIT-licensed and battle-tested via benchmarks; worth a spin for niche needs.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.