Many apps have fundamental accessibility points, like lacking labels or low distinction. Automated instruments will help app builders catch fundamental points, however could be laborious to run or require writing devoted checks. On this work, we developed a system to generate accessibility reviews from cell apps by a collaborative course of with accessibility stakeholders at Apple. Our technique combines diverse knowledge assortment strategies (e.g., app crawling, guide recording) with an present accessibility scanner. Many such scanners are based mostly on single-screen scanning, and a key drawback in entire app accessibility reporting is to successfully de-duplicate and summarize points collected throughout an app. To this finish, we developed a display grouping mannequin with 96.9% accuracy (88.8% F1-score) and UI aspect matching heuristics with 97% accuracy (98.2% F1-score). We mix these applied sciences in a system to report and summarize distinctive points throughout an app, and evaluated our system with 18 accessibility-focused engineers and testers. The system can improve their present accessibility testing toolkit and tackle key limitations in present accessibility scanning instruments.