Conhacks 2026 · iOS · Computer Vision

Reach into
reality.

Gestura turns your iPhone's camera into a workspace. Point at the world. Tap, pinch, wave, twist — and watch objects on screen respond to your hand.

Try the gestures How it works
9:41
∠ 0°
What it is

A camera
that listens to your hands.

Gestura is an iOS app that recognizes four intuitive hand gestures and applies them directly to virtual objects in your camera view — no controllers, no buttons, no menus.

See

The live camera feed becomes your canvas. Whatever's in frame is yours to work with.

Touch

Your index finger is the pointer. A single tap selects, a second tap releases.

Shape

Resize, rotate, or remove objects with the same gestures you already use every day.

The Gestures

Four gestures.
Endless control.

Every interaction starts with a tap. Try each one — the demos respond to you.

01 · Select

Tap to select.

Reach out and touch any object on screen with your index finger. Gestura highlights it in blue — that's your active object.

Tap it again to release. Tap a different object to switch focus.

Index-finger tap
Try it
Click the shape
02 · Resize

Pinch to scale.

Spread your index finger and thumb apart and the object grows with them. Bring them in and it shrinks.

The change is continuous — pixel-perfect, gesture-driven.

Pinch out / in
Try it
100%
03 · Remove

Wave it away.

One swipe of your hand to the right and the selected object disappears from the scene.

Clean canvas. Instant reset. Move on.

Wave right
Try it
04 · Rotate

Twist with a fingertip.

Move your index finger in a clockwise circle. The object rotates with it — naturally, smoothly, in real time.

The exact angle reads back on screen as you go.

Clockwise twist
Try it
∠ 0°

Selection comes first.

Resize, remove, and rotate are intentionally locked behind a tap. Nothing changes by accident — every transformation is something you chose.

Under the hood

How it works.

Three layers, running together at every frame of the camera feed — entirely on-device.

Capture

The iPhone camera streams a live feed at 60 fps. Every frame is passed straight into the vision pipeline.

Track

A hand-pose model identifies 21 landmarks per hand — fingertips, joints, palm — and classifies the gesture in real time.

Respond

The selected object reads the gesture's parameters — distance, angle, direction — and updates its scale, rotation, or visibility instantly.

Built with

An iOS-native stack.

Designed to run entirely on-device. No cloud, no latency, no shared frames.

SwiftUI

Interface and overlays

ARKit

World and object placement

Vision

Hand-pose detection

Core ML

Gesture classification