Segment Anything vs Flow

Enjoy state-of-the-art image segmentation on your iPhone or iPad. All processing is done locally on your iphone or ipad, everything works without network. We've optimized it for smooth, reliable performance on your device. The implementation of Segment Anything is based on the open source SAM(Segment Anything Model). In Pro mode, users will be able to unlock the following features: 1. Instead of being limited to adding only 3 "Add mask" areas, users can add up to 50 in the Pro mode. 2. The "Remove area" feature becomes available, providing users more flexibility and control over the image cutout process. 3. Users have the ability to download their cutout images, perfect for saving, sharing, or using in other applications.

Flow turns your iPhone and iPad into a powerful sensor hub. Build visual data pipelines by connecting nodes on a canvas — stream camera, motion, and AR data to any device over OSC, UDP, or TCP. VISUAL NODE EDITOR Drag, connect, and configure nodes on an interactive canvas. No coding required. Build complex sensor routing in minutes with an intuitive drag-and-connect interface inspired by tools like TouchDesigner and Max/MSP. 50+ NODES IN 3 CATEGORIES Sources — Camera & Vision (body pose, hand tracking, face landmarks), Sensors (accelerometer, gyroscope, magnetometer, barometer, ARKit), Generators (signal, noise, counter, timer, clock, array sequencer), and Data (API request with JSON extraction). Processing — Math operations, trigonometry, normalize, quantize, smooth filters, map range, waveform modulation, oscillators, logic gates, comparators, thresholds, triggers, delay, sample & hold, looper, vector math, string formatting, and JSON building. Destinations — TCP client & server, UDP sender, OSC sender. Plus multi-device nodes that broadcast to sequential IP addresses with per-device value offsets — perfect for LED installations, coordinated displays, and multi-device performances. REAL-TIME LIVE DATA Monitor every node in real time. Source nodes show output data, processing nodes show both input and processed output, destination nodes show what they receive. Human-readable formatting for all data types — JSON pretty-printing, pose joint details with confidence bars, hex blob previews, matrix grids, and sparkline history charts. MULTI-DEVICE ORCHESTRATION Control multiple devices simultaneously. Set a base IP and device count — Flow sends to each device with optional per-device value offsets. Create wave effects, staggered animations, or synchronized installations across OSC, UDP, or TCP. API INTEGRATION Fetch data from any REST API with configurable HTTP method, headers, and request body. Extract values from JSON responses using dot-notation key paths. Auto-fetch on intervals or trigger manually. 30+ READY-MADE TEMPLATES Start instantly with pre-built workflows for hand tracking, body pose, face landmarks, accelerometer streaming, ARKit world tracking, noise generation, signal processing chains, multi-device broadcasting, and more. CAMERA PREVIEW Live camera feed appears automatically on the canvas when running vision pipelines. See your pose tracking, hand detection, or face landmarks overlaid in real time. SAVE & REUSE Save workflows and destination presets. Quickly switch between network configurations. Load saved workflows from the template picker. BUILT FOR CREATIVE TECHNOLOGISTS Flow is designed for artists, VJs, installation designers, creative coders, and anyone who needs to get sensor data from iOS to other software or hardware. Works seamlessly with TouchDesigner, Max/MSP, Processing, openFrameworks, Unity, Unreal Engine, Resolume, and any OSC/UDP/TCP-compatible tool. TECHNICAL HIGHLIGHTS • iOS 17+ with SwiftUI and Swift Concurrency • 60fps pipeline engine with actor-based architecture • Network.framework for low-latency TCP/UDP • Full OSC 1.0 protocol support • Vision framework for real-time pose estimation • CoreMotion for all device sensors • ARKit for world and face tracking