Apps where the camera is a primary interaction surface, document scanning, barcode workflows, inspection capture, and media creation, built with react-native-vision-camera for full camera API access including custom frame processors that run native code on every camera frame without JavaScript bridge overhead. Document scanning with perspective correction: vision-camera-code-scanner for real-time barcode and QR detection; custom frame processor running OpenCV-based edge detection to identify document boundaries and apply perspective transform, producing a flat corrected scan from a photo taken at an angle, the same result a hardware document scanner produces but from a phone camera. Inspection and field capture workflows: photo capture tied to a structured form (the technician photographs the defect, then fills in the defect type, severity, and location fields in the same screen); batch capture mode for multi-point inspections; automatic metadata attachment (GPS coordinates, device timestamp, user ID) to each photo before upload. On-device ML inference using TensorFlow Lite or Core ML via react-native-fast-tflite for real-time image classification, object identification, defect detection, or document type classification running locally without a network round-trip. Video recording with quality configuration (resolution, bitrate, frame rate) and upload via multipart streaming to S3 or a CDN origin, with upload progress displayed and resume-on-reconnect for large files on mobile networks. AR features using Expo Camera or ViroReact for ARKit (iOS) and ARCore (Android) integrations where the requirements stay within the cross-platform layer's capabilities.