TypeSafe MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@TypeSafe MCP ServerClassify this ticket: Payment gateway timeout on checkout."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
TypeSafe MCP Server
Model Context Protocol (MCP) 服务器,为 Claude 提供 TypeSafe Jev 和 Codex 模型的结构化判断能力。
功能特性
1. 通用结构化判断 (typesafe_classify)
支持三种问题类型:
choice - 单选/多选分类
score - 数值评分(1-10)
noul - 布尔判断(是/否概率)
2. 工单智能分类 (typesafe_ticket_classify)
自动分析工单并输出:
类型(bug/性能/需求/咨询)
优先级(P0-P3)
负责部门(后端/前端/运维/产品/客服)
用户情绪(平静/焦急/愤怒)
影响范围评分(1-10)
3. 内容审核 (typesafe_content_moderate)
检测违规内容:
暴力/色情/仇恨言论/垃圾信息
风险评分(1-10)
审核建议(通过/人工审核/拒绝)
Related MCP server: OCR MCP Server
安装
1. 安装依赖
pip install -r requirements.txt2. 获取 TypeSafe API Key
访问 TypeSafe 注册并获取 API Key。
3. 配置 Claude Code
编辑 ~/.claude.json(或项目的 .claude/settings.json),添加:
{
"mcpServers": {
"typesafe": {
"command": "python",
"args": ["/path/to/typesafe-mcp/server.py"],
"env": {
"TYPESAFE_API_KEY": "your-api-key-here"
}
}
}
}注意:
将
/path/to/typesafe-mcp/server.py替换为实际路径Windows 用户可能需要使用 Python 的完整路径(如
C:\\Python313\\python.exe)
4. 重启 Claude Code
# 或在 Claude Code 中执行
/reload使用示例
通用分类
# 在 Claude 对话中:
"使用 typesafe_classify 工具,判断这段文本的情感:'这个产品太棒了!'"
# 问题定义:
{
"sentiment": {
"type": "choice",
"criteria": {
"positive": "积极正面",
"neutral": "中性客观",
"negative": "消极负面"
}
},
"confidence": {
"type": "score",
"criteria": ["判断置信度", "1=不确定", "10=非常确定"],
"score": {"min": 1, "max": 10}
}
}工单分类
# 在 Claude 对话中:
"使用 typesafe_ticket_classify 工具,分析这个工单:
标题:登录页面加载超过10秒
内容:用户反馈从昨晚开始,登录页面变得特别慢,有时甚至超时"内容审核
# 在 Claude 对话中:
"使用 typesafe_content_moderate 工具,审核这段内容:'...'"环境变量
变量名 | 必需 | 默认值 | 说明 |
| ✓ | - | TypeSafe API 密钥 |
| ✗ |
| API 端点 |
开发
项目结构
typesafe-mcp/
├── server.py # MCP 服务器主文件
├── requirements.txt # Python 依赖
└── README.md # 本文档自定义工具
可以在 server.py 中添加新的工具函数:
types.Tool(
name="your_tool_name",
description="工具描述",
inputSchema={
"type": "object",
"properties": {
"param": {"type": "string", "description": "参数说明"}
},
"required": ["param"]
}
)故障排查
MCP 服务器未加载
检查
~/.claude.json配置格式确认 Python 路径和脚本路径正确
查看 Claude Code 日志:
~/.claude/logs/
API 调用失败
验证
TYPESAFE_API_KEY是否正确检查网络连接
确认 API Key 额度是否充足
工具未显示
重启 Claude Code 或执行 /reload
许可证
MIT License
相关链接
This server cannot be deployed
Maintenance
Related MCP Connectors
Sentiment, toxicity, entity extraction, PII, translation, summary, QA, fraud scoring, safety audit.
Turn messy text into strict JSON schemas agents can trust (invoice, receipt, contact, resume).
- SnipgetOAuthai.snipget
300+ deterministic data utilities for AI agents: validate, normalize, parse, match, redact.
A paid remote MCP for Pydantic AI structured output, built to return verdicts, receipts, usage logs,
Related MCP Servers
- AlicenseBqualityDmaintenanceProvides Claude with 44 tools for confidence gating, typed outputs, hallucination detection, and constraint enforcement during conversations.511MIT
- AlicenseNot gradedqualityCmaintenanceEnables document text recognition and extraction from images and PDFs using Claude Vision, including support for scanned documents and structured output, without requiring local OCR engines.MIT
- AlicenseAqualityBmaintenanceEnables typed, calibrated judgment calls through classify, score, check, and batched ask tools, each returning full probability distributions for programmatic decisions.5981 npm7MIT
- AlicenseAqualityCmaintenanceEnables agents to call typed code-review and content-moderation decision tools, returning structured verdicts, probabilities, and confidence-gated actions.2MIT