Beyond Vision turns your iPhone camera and photos into live haptic sensations. Glide your finger. Feel objects pulse. Hear their names.
A multi-sensory experience built entirely on native Apple frameworks — no internet required.
Import any photo and drag your finger freely. Objects become invisible haptic hotspots — stronger as you centre on them, fading at the edges.
Point your camera at the world and touch what's around you in real time. Detect objects, animals, text, and general areas of interest instantly.
The moment your finger contacts an object, its name is spoken aloud. Read signs, labels, and text by simply touching them on screen.
YOLO detection, animal recognition, saliency mapping, fine-grained classification, and OCR — all running on-device. No data leaves your phone.
Adjust vibration intensity and sharpness to match your personal sensitivity. Your experience, your feel.
Designed for users who are blind or have low vision. No visual attention required — every interaction is driven by touch and audio.
A real-time pipeline that translates vision into vibration in milliseconds.
The camera streams live frames, or you load a photo from your library. Either way, the pipeline starts immediately.
YOLO, Apple's animal recogniser, and saliency mapping run in parallel to find every object and point of interest in the frame.
Detected bounding boxes are mapped onto the screen as invisible haptic zones, precisely aligned with what the camera sees.
Your touch distance from each object's centre modulates haptic intensity in real time. Find it, feel it pulse, hear its name.
Having trouble or have a question? Reach out through any of the channels below.
Send us a detailed message and we'll get back to you as soon as possible.
swastikmandal.2416@gmail.com →Found something that isn't working right? Open an issue on GitHub with steps to reproduce.
Open an issue →Have an idea that would make Beyond Vision better? We'd love to hear it.
Suggest a feature →Beyond Vision is open source. Explore the codebase, contribute, or fork it for your own project.
View on GitHub →