Overview

Claude Compliance gives your organization visibility into Claude.ai usage by connecting to Anthropic's Compliance API. Once you register a Compliance API key, the platform continuously fetches your organization's activity data and runs DLP scanning on all user inputs - surfacing findings directly in the dashboard.

What it provides

Capability	Description
Activity feed	Full log of user actions: logins, chat interactions, file uploads, and administrative events
Organizations & users	List of all organizations under your parent org and their member users
Chats & files	Chat message content and file attachments from user sessions
Projects	Project names, descriptions, instructions, and attached documents
DLP findings	Automatic scanning of all user inputs for sensitive data - PII, credentials, financial data, and more

How it works

1. Register a Compliance API key - your sk-ant-api01-… key is validated and stored securely. The plaintext key is never exposed after registration.
2. Data is synced automatically - organizations, users, chats, projects, and activities are fetched from the Compliance API on a regular schedule.
3. Inputs are scanned for DLP - user message text, file attachments, and project content are scanned and classified by severity.
4. Findings are surfaced - DLP results are available through the dashboard, filtered by time, severity, user, and more.

DLP severity levels

Level	Triggers
CRITICAL	SSN patterns, credit card numbers, API/secret keys, private key blocks
HIGH	Passwords shared in chat, bearer tokens
MEDIUM	Email addresses, phone numbers
LOW	IP addresses
NONE	No signals detected

Data sources scanned

Source	What is scanned
chat_message	User turn text in a chat conversation
chat_file	Text files attached to chat messages
project_description	The description field of a project
project_instruction	The system prompt / instructions field of a project
project_file	Text files attached to a project

---

Next: See Key Management to learn how to register, monitor, and revoke Compliance API keys.