pii-mcp
MCP server for anonymizing and deanonymizing PII through the Pseudora API, enabling safe sharing of sensitive text with AI assistants.
README
pii-mcp
MCP (Model Context Protocol) server that exposes pii-protect as native tools for Claude Desktop, Cursor, and any MCP-compatible client.
Tools exposed
| Tool | Description |
|---|---|
anonymize_text |
Anonymize PII, returns anonymized text + context_id |
deanonymize_text |
Restore original using context_id |
list_pii_types |
List all detectable PII types |
Connect it to your Pseudora account
- Sign in at pseudora.cloud and open
Cloud → Customer keys (
/cloud/customer-keys). - Create a key. It is displayed once — copy it straight away.
- Point the server at the hosted API:
pip install -r requirements.txt
cp .env.example .env
PII_API_URL=https://pseudora.cloud
PII_API_KEY=pk_live_…
The key travels as an X-Api-Key header. Your Pseudora plan decides the rate
limit and the monthly credit pool, both shared with every other client on the
same account — the browser extension, the VS Code extension and this server all
draw from the same pool. A key belonging to an account with no active
subscription is refused with 402 subscription_expired.
To run against a self-hosted engine instead, set PII_API_URL to your own
instance (for example http://localhost:15500) and use a service or admin key.
One-shot setup
bash setup.sh
It prompts for the key and registers the server with Claude Desktop.
Claude Desktop configuration
Edit ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"pii-protect": {
"command": "python",
"args": ["/Users/you/Progetti/pii/pii-mcp/server.py"],
"env": {
"PII_API_URL": "https://pseudora.cloud",
"PII_API_KEY": "your-customer-api-key"
}
}
}
}
Restart Claude Desktop. The tools appear automatically in the tool menu.
Cursor configuration
Edit .cursor/mcp.json in your project (or global ~/.cursor/mcp.json):
{
"mcpServers": {
"pii-protect": {
"command": "python",
"args": ["/Users/you/Progetti/pii/pii-mcp/server.py"],
"env": {
"PII_API_URL": "https://pseudora.cloud",
"PII_API_KEY": "your-customer-api-key"
}
}
}
}
Without the mcp SDK
server.py includes a pure JSON-RPC 2.0 fallback that works without the mcp package.
The mcp SDK is optional but recommended for full protocol compatibility.
# With SDK (recommended)
pip install mcp
# Without SDK (fallback mode)
pip install httpx python-dotenv
Usage example (Claude Desktop)
After connecting, ask Claude:
"Anonymize this text before I share it with you: Mario Rossi, CF RSSMRA80A01H501U, tel 333-1234567"
Claude will call anonymize_text and work only with the anonymized version.
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。