Virdio
Designing an AR Fitness Platform Across Every Screen
TL;DR
Home fitness users wanted engaging classes without expensive equipment, but no product delivered AR workouts across all devices.
Virdio's AR fitness tech could run on anything with a screen and camera, my role was to adapt the technology into a cross platform AR fitness subscription service.
I designed critical features across platforms: room setup, HUD, scheduling, onboarding, and a cross-platform design system.
Virdio successfully launched on all platforms, partnering with gyms and trainers to host hundreds of classes.
Expensivehardwarewasthebarriertoengaginghomefitness
“How might we make AR workout overlays usable, readable, and motivating across wildly different screen sizes, processing capabilities, and home environments?”
In 2021, millions of people were working out at home, but the most engaging options required expensive, space-consuming equipment. Peloton needed a $1,500 bike. Mirror needed a $1,500 screen. Everyone else was stuck with pre-recorded YouTube videos and zero performance tracking.
Virdio’s leadership had machine vision technology that could detect body poses through a standard camera and simulate exercise equipment using AR. The opportunity was enormous: deliver the engagement of a Peloton class to anyone with a laptop or phone, no hardware required.
Peloton
$1,500 bike
Mirror
$1,500 screen
YouTube
Free / No tracking
Virdio
Any camera
An early-stage startup, one designer, five platforms
Virdio had been licensing its machine vision technology to gyms for remote AR classes. The next step was a direct-to-consumer subscription app, like Peloton but accessible from any device. I was brought on as the sole full-time product designer to take this from zero to launch.
The product needed to ship on iOS, Android, web, desktop (Mac and Windows), Apple Watch, and smart TVs. The engineering team was fully remote and 12 hours ahead of me, which meant nearly all dev collaboration was asynchronous. There was no existing design system, no prior consumer-facing product, and the timeline was aggressive.
Desktop is where the AR magic happens. Mobile is the browsing front door.
Themostaccessibledevicedeliveredtheworstexperience
I started testing prototypes with our internal advisory board of fitness trainers and physicians. What emerged was a turning point: the mobile phone, the easiest way for users to access the app, was inherently the worst platform for the core AR workout experience.
On a small screen, AR overlays competed with the video feed for visibility. The pose detection needed distance from the camera, but users placed phones close or on unstable surfaces, leading to inconsistent tracking or difficult visibility. If mobile was the front door for most users, and mobile delivered the lowest-quality version of our differentiating feature, we risked first impressions that undermined the entire value proposition.
Mobile
Most accessible
Small screen, unstable placement, AR overlays compete for visibility.
Desktop
Best AR quality
Larger screen, stable camera, overlays complement the experience.
More accessible devices tended to deliver lower-quality AR; desktop was the best place for the core workout experience.
Platform strategy: launching everywhere vs. focusing on the best experience
This insight forced a critical strategic conversation. I advocated for a desktop-first approach: focus our limited resources on making the larger-screen experience exceptional. The CEO was adamant about launching on all platforms simultaneously. Multi-platform availability was a competitive differentiator, and mobile was the most accessible entry point for browsing and booking classes.
Option A
Desktop/Laptop First
Focus on larger screens where AR overlays were most usable, then expand to mobile later. Ship a polished hero experience faster with our small team.
Option B
All Platforms Simultaneously
Launch everywhere at once to maximize accessibility and differentiate from hardware-locked competitors. Mobile is the most accessible entry point.
What we chose
We shipped on all platforms but designed the UX to guide users toward the desktop/laptop experience as the recommended way to attend AR classes. Mobile retained full functionality but onboarding, class setup, and messaging nudged users toward larger screens.
What I gave up
The ability to ship a deeply polished single-platform experience. I spread my effort across five platforms, which meant every surface got less iteration time than I wanted.
ARworkoutclassesanyonecouldsetupintheirlivingroom
I designed an end-to-end experience that made AR fitness accessible regardless of device or home environment: from browsing and booking to live workouts with AI-powered pose detection and AR equipment overlays, all through an existing camera.
Room Setup & Calibration
Users needed to calibrate their camera and define their play space, which sounds technical. I designed a visual experience to make it approachable. For camera alignment, I created a visual guide that asked users to center themselves on screen with clear tilt indicators.
For room calibration, the system placed virtual cones and users simply walked to the corners of their space. The backend detection handled the rest. The entire setup culminated in a satisfying green checkmark confirmation.
Start Setup
Approach Cone
Capturing
Cone Complete
Next Cone
Capturing
Setup Complete
Mobile version of the room calibration had more prominent text and larger UI elements so the user could spot them better on a smaller screen.
What if setup fails?
This flow covers what happens if calibration fails during a class or if a user skips setup: how the product recovers, what users see next, and when they can retry or continue with degraded AR.
No-blame UI: “Did you step out of frame?” frames it as a system state, not a user error. Options let users ignore or get reminded later.
Interactive
Try It: Calibrate Your Space
Place 2 cones at opposite corners to define your workout zone. Drag to move, double-click to remove.
Live Workout HUD
The core of the product used machine vision to read body poses in real time, counting actions like punches, squats, and jumps. AR artifacts served as both affordances showing users how to perform exercises and as hit boxes for the system to register completed reps.
I designed platform-specific HUD adaptations. Desktop took advantage of landscape orientation to show more metrics, class info, and participant data simultaneously. Mobile required a dynamic HUD that collapsed and expanded contextually, preserving maximum visibility of the actual workout content on a small screen.
The desktop AR experience at its best: full metrics, timer, and AR equipment track visible at once.
Desktop shows everything at once. Mobile collapses to preserve workout visibility.
The AR Exercise Library
Scheduling & Design System
I built the design system from the ground up to maintain consistency across all five platforms. It governed color, typography, button styles, and component behavior. A key design decision was using light mode for browsing and discovery surfaces and dark mode for anything related to attending a class, creating a clear psychological shift when users entered the workout experience.
Desktop exposes filters inline for power browsing. Mobile surfaces categories as compact chips.
“My Schedule” connects browsing to action: calendar, booked classes, and a prominent Join Session CTA.
Role&Impact
What I Owned
- •Sole designer from concept to launch-ready deliverables across iOS, Android, web, desktop, Apple Watch, and smart TV.
- •Built the entire design system from the ground up: color, typography, components, and platform-specific adaptations.
- •Influenced product strategy: led the platform prioritization debate and shaped the desktop-first UX compromise.
- •Created detailed interaction and animation specs for async engineering handoff across a 12-hour timezone gap, critical for AR components that static mockups couldn't convey.
Impact
The app launched across all planned platforms in mid-2022. I was let go about a month before launch during team reductions, just after completing the final design iterations. The app received some positive reviews but did not achieve significant consumer adoption. The consumer home fitness market was cooling from its pandemic peak by the time we shipped.
If I could do it over, I would focus on making one platform exceptional with a limited, curated class library, and use that to wow users before expanding.