SDKs collect runtime signals when your app runs:

• Performance: Core Web Vitals and timings
• Errors: exceptions and stack traces
• User behavior: sessions, page views, key interactions
• Security signals: basic heuristic detections (plan/features dependent)

• Deduplication: group similar errors by signature
• Aggregation: compute trends over time
• Classification: severity and impact tagging

The Agent ranks SDK telemetry plus scanner output into issues (errors, vulns, slow LCP per route, instability). The dashboard shows them in Issue List (SDK events) on the Security tab and in Overview/Summary counts. TestingAI traffic is filtered out of Security.

Fixes: manual View Fix → Commit Correction or Agent Autonomous.

TestingAI runs headless browsers (Playwright/Chromium) on the Sublyzer backend to generate synthetic navigation, videos, and logs.

See: TestingAI Guide

Retention varies by plan and policy. You can export or delete data from Settings.