SkyFiCam vs Dual Camera

SKYFI turns your camera into a cinematic sci-fi viewfinder. Point it at any open sky and an on-device neural network identifies the sky region in real time, replacing it with stunning procedural effects while leaving buildings, trees, and people untouched. EVERYTHING RUNS LOCALLY Every frame stays on your device. SkyFi uses Apple's Core ML, Vision, and Metal frameworks to segment the sky, render effects, and record video — all on the Neural Engine and GPU. No frames are uploaded. No accounts. No tracking. FOUR LENSES • Original Feed — clean, unfiltered live camera. • Chaotic Sky — three glowing UAPs drift across the sky, ringed by FLIR targeting brackets. Inspired by the publicly released Navy GIMBAL / Tic-Tac / GOFAST footage. • UFO Monolith — a real-time raymarched obsidian slab floats above the city, lit by a cyan fresnel edge glow. Foreground objects occlude the lower edges automatically. • Debug Sky Tint — visualizes the sky-mask in red so you can see exactly what the AI sees. FLIR ↔ COLOR TOGGLE Tap the chip in the HUD to flip the entire feed between full color and Navy-thermal-style black-and-white. Pair it with the Chaotic Sky lens for the iconic Pentagon-footage look, or with the Monolith lens for a 2001-style cinematic monochrome. LIVE VIDEO RECORDING Tap the shutter to record at 60fps with the lens effect and FLIR filter baked into the saved file. Videos save to your Photos library and match the on-screen aspect (9:16 portrait or 16:9 landscape). SMART AND HONEST • If you're indoors or pointing at a wall, SkyFi tells you ("NO SKY DETECTED") and falls back to the unmodified camera instead of faking effects. • Live telemetry shows segmentation latency, frame rate, and processor usage so you can see the AI working. • Thermal-aware: under heavy load the segmenter throttles automatically to keep the device cool and the preview smooth. WHAT MAKES IT DIFFERENT SkyFi isn't a static-filter app. The neural network re-segments your sky every few frames using Apple's DETR Resnet-50 semantic segmentation model. The lens effects are procedurally generated with Metal SDF shaders — there are no canned video assets, every frame is computed on the fly. REQUIREMENTS iPhone with iOS 16 or later. Apple Neural Engine recommended (iPhone XS or newer) for best performance. PRIVACY Camera access is required for the live preview. Photo Library "add only" access is required to save recordings. No data is collected, transmitted, or shared. Open the camera privacy section in iOS Settings to revoke access at any time.

DualCamera is a video recorder built for two-view capture and multi-platform content creation. Record with the front and rear cameras together to capture the scene and your reaction, or create 16:9 landscape and 9:16 portrait videos from a single take. THREE RECORDING MODES • Standard Record everyday video with the front or rear camera. • Front & Rear Capture both cameras at the same time in a picture-in-picture layout. Move the inset, tap to swap the main view, choose Rectangle, Capsule, Square, or Circle, customize the border color and width, or save both camera sources separately for editing. • Wide & Tall Record once and save both 16:9 landscape and 9:16 portrait versions, ready for horizontal and vertical platforms. MORE CAMERA CONTROLS • Full Screen, 9:16, 3:4, 1:1, 4:3, and 16:9 aspect ratios • 720p, 1080p, and 4K resolution • 24, 30, and 60 fps • Video stabilization, flash, grid, and smooth zoom • Mono or stereo audio • Optional location metadata • Optional separate front and rear source clips Whether you are making vlogs, reaction videos, interviews, tutorials, product demos, travel stories, or both landscape and portrait edits, DualCamera helps you get more useful footage from every take.