BeLikeNative Grammar Server
L1-aware grammar, style, translation & tone tools with 70 local rules. Zero API keys needed.
README
BeLikeNative MCP Server
A Model Context Protocol (MCP) server that provides grammar checking, writing improvement, translation, and tone adjustment tools to AI clients like Claude Desktop, ChatGPT, Cursor, and others.
No API key required. Grammar and style checks use a local rule-based engine. Translation and tone adjustment return structured prompts for the host AI to process.
Tools
| Tool | Description | Processing |
|---|---|---|
check_grammar |
Check grammar, spelling, and punctuation with L1-aware explanations | Local rule-based (50+ regex rules) |
improve_writing |
Analyze text for style, wordiness, passive voice, sentence length | Local rule-based + style guidelines |
translate |
Translate text between languages with natural, fluent output | Returns prompt for host AI |
adjust_tone |
Adjust text tone (formal, casual, professional, diplomatic, etc.) | Returns prompt for host AI |
Prerequisites
- Node.js 18+
That's it. No API keys, no environment variables, no external services.
Installation
cd mcp-server
pnpm install
Running Standalone
pnpm start
The server communicates via stdio (stdin/stdout). It is designed to be launched by an MCP client, not run interactively.
MCP Client Configuration
Claude Desktop
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"belikenative": {
"command": "node",
"args": ["/absolute/path/to/mcp-server/src/index.mjs"]
}
}
}
Claude Code
Add to your Claude Code MCP settings:
{
"mcpServers": {
"belikenative": {
"command": "node",
"args": ["/absolute/path/to/mcp-server/src/index.mjs"]
}
}
}
Cursor
Add to your Cursor MCP config (.cursor/mcp.json in your project or ~/.cursor/mcp.json globally):
{
"mcpServers": {
"belikenative": {
"command": "node",
"args": ["/absolute/path/to/mcp-server/src/index.mjs"]
}
}
}
Windsurf / Other MCP Clients
Any MCP client that supports stdio transport can use this server. Point it at node src/index.mjs. No environment variables needed.
Tool Schemas
check_grammar
{
"text": "string (required, max 6000 chars)",
"language": "string (optional, default 'en') -- writer's native language for L1-tailored explanations"
}
improve_writing
{
"text": "string (required, max 6000 chars)",
"style": "enum: academic | business | creative | technical | simple | concise (optional, default 'business')"
}
translate
{
"text": "string (required, max 6000 chars)",
"source_language": "string (required) -- e.g. 'en', 'English', 'fr'",
"target_language": "string (required) -- e.g. 'es', 'Spanish', 'de'"
}
adjust_tone
{
"text": "string (required, max 6000 chars)",
"tone": "enum: formal | casual | friendly | professional | persuasive | confident | empathetic | diplomatic (required)"
}
Architecture
src/
index.mjs -- MCP server entry point (stdio transport, tool registration)
tools.mjs -- Tool definitions (JSON schemas) and handler functions
rules.mjs -- Local grammar rules engine (50+ regex patterns, style analyzer)
- Transport: stdio (standard for MCP)
- Grammar/Style: Local rule-based engine (no external API calls)
- Translate/Tone: Returns structured prompts for the host AI client to process
- Logging: All logs go to stderr (stdout is reserved for MCP protocol)
- Error handling: Never crashes -- all errors return structured MCP error responses
How It Works
The key insight: MCP tools are called by AI clients (Claude Desktop, Cursor, etc.) that already have AI built in. There is no need for the MCP server to make its own API calls.
- check_grammar and improve_writing use 50+ regex-based rules to detect grammar errors, spelling mistakes, style issues, passive voice, and sentence length problems. Results are deterministic and instant.
- translate and adjust_tone genuinely require AI intelligence, so they return structured prompts with guidelines that the host AI processes directly. This is faster, cheaper, and more reliable than a double API call.
Code Quality
This server follows NASA Power of 10 rules:
- All functions under 60 lines
- Minimum 2 assertions per function
- All loops have fixed upper bounds
- No global mutable state (constants are frozen)
- Every return value is checked
- Zero warnings
BeLikeNative Developer Tools
This tool is part of the BeLikeNative ecosystem — AI-powered writing tools for non-native English speakers.
| Tool | Type | Description |
|---|---|---|
| Grammar Check | GitHub Action | PR grammar checker with 60 rules and L1-aware insights |
| Writing Assistant | GitHub Action | Writing quality analysis: readability, structure, clarity |
| i18n Checker | GitHub Action | Find hardcoded strings that need internationalization |
| Commit Lint | GitHub Action | Commit message grammar, format & clarity checker |
| Website Grader | Web Tool | Free website performance grader |
BeLikeNative Chrome Extension — AI writing assistant for 100+ languages, 15 tones, 15 styles. 10,000+ users, 4.6★ rating.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。