BLOG-3313 1

Cloudflare's CASB now consumes the Claude Compliance API, giving security and compliance teams visibility into Claude usage inside the Cloudflare dashboard with no endpoint agents involved. The integration extends Cloudflare's existing AI governance coverage to one of the most widely used AI tools in enterprise stacks.

Why sanctioned AI tools need separate coverage

Most organizations today have only partial visibility: unauthorized AI tools can be blocked at the network layer, but activity inside sanctioned ones stays hidden. Claude and similar tools differ from traditional SaaS in ways that conventional security tooling does not address. They are conversational, persistent, and embedded in workflows through APIs and agent frameworks, and they generate content rather than just storing it. A user may paste customer data into a prompt, a developer may share an API key and leave it unrotated, or a model response may carry company secrets. None of these are reliably detectable by network controls alone.

Cloudflare's answer spans the full workflow:

  • AI Gateway provides observability into requests, token spend, and model performance for traffic between applications and providers such as Anthropic, with controls for rate limits, response caching, and routing.
  • Gateway and Data Loss Prevention inspect AI traffic and block prompts containing customer PII or confidential material before they reach the model.
  • Access with MCP server portals centralizes agent connections to corporate tools behind one protected endpoint, with audit logging on every request.
  • CASB covers data at rest inside Claude, scanning for misconfigurations and sensitive data without inline inspection.

Because these services run side by side on the same metal, they are composable and programmable, and traffic does not hairpin through multiple vendors or clouds to be secured.

How the CASB integration works

Cloudflare CASB connects to third-party SaaS applications through lightweight API integrations and scans them for misconfigurations, improper sharing, and related risks. The Claude Compliance API, built by Anthropic, gives enterprises programmatic access to security-relevant data about their Claude organizations, workspaces, and usage. Cloudflare CASB consumes that endpoint and produces findings through the detection and remediation workflows Cloudflare One customers already use.

The integration supports the following asset types from launch:

  • Projects — projects shared with the whole organization or with a subset of users and groups.
  • Project attachments — files and documents in projects that violate DLP policies.
  • Chat files — user-uploaded and provider-generated files that violate DLP policies.
  • Chat messages — user prompts and provider responses that violate DLP policies.
  • Artifacts — provider-generated documents and files that violate DLP policies.

Findings land in the dashboard alongside posture and content findings from other SaaS applications, grouped by category and ordered by severity level. Teams triage and remediate them with the same workflows used for Microsoft 365, Google Workspace, or Salesforce.

Coverage across Claude products

For Claude Enterprise, CASB surfaces compliance data including organizations, projects, chats, and roles, and pulls conversation content through dedicated read-only endpoints to guard against data loss. For Claude Platform, CASB surfaces member and workspace changes, API key creation, and file create or download events, with Activity Feed support planned.

CASB findings can be turned into enforcement quickly. A finding such as a user uploading sensitive files can become a Gateway policy in minutes: block uploads to Claude for specific users, restrict access to the application entirely, or limit functionality until the issue is resolved. This links CASB detection to Cloudflare's existing inline policy engine.

Enabling the integration

  1. Ensure you have a Claude Enterprise account.
  2. Request Compliance API access from Claude for your organization.
  3. In the Cloudflare dashboard, go to Zero Trust > Integrations > Cloud & SaaS.
  4. Select Add Integration > Anthropic and enter your Compliance API key.
  5. Configure DLP profiles if uploaded files should be scanned for sensitive data.

Scanning starts immediately, and findings appear in the dashboard within minutes. New Cloudflare customers can sign up and begin with their first two integrations free; existing customers can enable the integration directly in the dashboard.

Roadmap

Cloudflare plans to keep expanding CASB coverage for AI tools as providers ship new enterprise security APIs. It is also deepening CASB integrations so customers can define custom findings and build workflows that remediate findings automatically. Documentation is available for developers, and a changelog subscription provides automatic updates.