Skip to main content
Choose the path that matches how you want to use DataFog.

Use DataFog MCP

Check local files before your agent opens them. Connect Claude Code, scan for personal data and credentials, and create protected copies.

Build with DataFog Core

Add detection and privacy transformations to your application with the Rust, Python, Node.js, or browser SDK.

Use MCP with your agent

Start with the MCP setup guide: install the server, choose allowed folders, and connect Claude Code. Then learn how to scan and transform files. DataFog MCP uses Core for detection. You do not need to write SDK code to use it.

Build with the Core SDK

Run the Core quickstart to detect and redact an email address in your runtime. For package choices, see installation. For the engine’s capabilities and operation model, see the Core overview. Already using the older datafog Python package? Follow the migration guide.

Explore the research

The local PII model research preview covers a separate experimental detector, its evaluation results, and its limitations.