com.anthropic.claude Security & Privacy Scorecard

Claude

What This Means for You

Your device information and a persistent device identifier are shared with Sift Science for fraud prevention and Segment for analytics, and some of this collection is not reflected in the App Store privacy label shown to you before installing the app.

Recommendation: Acceptable with Awareness

Secure encryption and strong authentication are present, but the app uses multiple analytics and fraud-prevention SDKs with incomplete privacy label disclosure. Review the privacy policy for full data handling details.

Best For: Users comfortable with standard app analytics and fraud-prevention telemetry who want a full-featured AI assistant

Key Findings

Data Security - 5 findings (2 low, 3 info)

Network Security - 0 findings

Code Safety - 5 findings (3 low, 2 info)

Privacy - 5 findings (1 medium, 1 low, 3 info)

Privacy Concerns

What Data is Collected

• Personal information: account details used for sign-in (Google, Apple)
• Device information: device model, iOS version, network type, carrier, battery level, screen dimensions, persistent device identifier
• Usage data: app interactions, session analytics, behavioral patterns
• Location data: general region (via Google Maps SDK)

Third-Party Data Sharing

The following third parties may receive your data:

• Sift Science - fraud and risk scoring using device telemetry and a persistent device identifier
• Segment - session analytics and behavioral usage tracking
• Datadog - app performance monitoring and crash reporting
• Sentry - error reporting
• Firebase (Google) - push notifications and crash reporting
• Google - authentication (Sign-In) and map rendering

Understanding the Scores

Security: 98/100
Privacy: 92/100

Security Breakdown

• Data Security: 99/100 - Conversation data is strongly protected using the highest-tier iOS storage protections; minor issues are limited to third-party SDKs
• Network Security: 100/100 - All production traffic uses encrypted connections; no insecure transmission of user data detected
• Code Safety: 98/100 - Strong device integrity verification and sandboxed web content rendering; staging infrastructure and a debug menu remain compiled into the production app

Privacy Breakdown

• Data Collection: 93/100 - No advertising tracking or cross-app tracking; device telemetry collected by the fraud-prevention SDK is not declared in the App Store privacy label
• Data Sharing: 92/100 - Data shared with analytics and fraud-prevention vendors; some sharing is not accurately reflected in privacy manifests
• User Control: 100/100 - Conversation history is stored in your own iCloud account, giving you full control over your data

Positive Security Features

• Sensitive account data stored with the highest iOS protection level, only accessible when the device is unlocked with a passcode
• Apple App Attest implemented so the server can verify that requests come from genuine, unmodified versions of the app on real Apple hardware
• No advertising tracking or cross-app data sharing; the advertising tracking permission prompt is not required
• AI-generated content rendered in a tightly sandboxed environment, preventing it from accessing device resources or other app data
• Database query access controls prevent unauthorized data mutations or schema changes
• Firebase Analytics explicitly disabled despite the Firebase SDK being present

Areas for Improvement

GDPR / CCPA Compliance

The app's privacy practices could be strengthened by:

1. Privacy Label Completeness
   Sift Science collects a persistent device identifier and detailed device telemetry for fraud scoring, but this collection is not declared in the App Store privacy nutrition label. Users have no visibility into this data use before installing the app.

2. SDK Manifest Accuracy
   The Segment analytics SDK privacy manifest contains a malformed entry for device identifier collection (empty purpose field) and an inaccurate data-use category. Accurate manifests give users and regulators a clearer picture of how data is used.

Security Enhancements

1. Staging Infrastructure Removal
   Development server addresses, a debug menu module, and staging server associations are compiled into the production app. While the debug menu is protected by a server-side flag, removing these from production builds reduces the available attack surface.

2. Web Content Bridge Origin Validation
   The communication channel between the app and sandboxed web content does not verify the source of incoming messages. Adding origin checks would further protect the native bridge from unexpected message sources.

Technical Context

App Type: AI assistant handling sensitive conversational data and personal information
Classes Analyzed: 21
Third-Party Services: 16
Context Tags: sensitive_data, location

---

About This Analysis

This security analysis was conducted by CITT (Can I Trust That), an independent security analysis service. We perform static code analysis on iOS applications to help users make informed decisions about app security and privacy.

App Details

Developer: Anthropic
Version: 1.260409.1 (Build 24361703904)
Analysis Date: 2026-04-17
Package: com.anthropic.claude

Analysis Limitations

• Static analysis only (code review without running the app)
• Based on IPA version 1.260409.1 (Build 24361703904) analyzed on 2026-04-17
• May not reflect server-side security controls
• Cannot detect all runtime behaviors