Product · Mobile app · Engine

Unio

Every fitness app we tried was built for one person, with sharing added afterwards. Two people training together either followed one plan that suited neither of them, or used two apps and lost the point. So we built a workout app that starts from the pair: two separate plans, generated for two different bodies, on one shared week.

RoleProduct design, engine, iOS app, backend
StackCapacitor iOS · Supabase · RevenueCat · Vercel
StatusLaunching on iPhone
Year2026
trainwithunio.com
The Unio website
Hover to scroll

A live capture of the Unio site. Hover to scroll the page, or open it for real ↗

The problem

Two people who want to train together are not the same size, do not have the same goal, are not free on the same days and do not own the same equipment. A single shared plan is wrong for at least one of them, usually both.

The workout was never really the hard part. Training together falls apart the first week one person skips, because nothing makes the other one visible. The apps that exist treat the second person as a follower on a feed, not as half of the plan.

What we built

  1. A workout engine of our own Every session is generated by an engine we wrote, from the person's goal, the days they can train, the kit they have and what they actually lifted last time. No AI model and no API call behind it, so there is no cost per workout and nobody's body measurements leave our infrastructure. The same inputs always give the same plan, which is what makes it testable: every change is swept and fuzzed across generated weeks before it ships.
  2. Two plans on one week Each person gets a plan built from their own numbers. Both sit on one calendar with one shared streak, so each can see whether the other showed up, without either one having to follow the other's programme.
  3. 278 exercises, drawn and animated in house Every move in the library has its own animated figure, drawn by us rather than licensed, and posed from the angle that shows the joint actually under load. Lifting, yoga, pilates, calisthenics, stretching and cardio, down to the machine each one is done on.
  4. Training together, live When your partner starts a session you can see it happening and send a short video into it while they are still mid set. Clips are watched once and then gone, so nothing piles up into a feed.
  5. Pairing that survives an install Invites are single use links that carry a six character code, so pairing still works when the other person has to download the app first. iOS does not otherwise let you carry a link through the App Store, and without this the invite is lost at the exact moment it matters.
  6. One subscription for two Apple in app purchase through RevenueCat, where one subscription unlocks both people in the pair. Priced for the couple, not the individual.
The hard part of training together was never the plan. It was both people still showing up in week three.

Where it landed

278Exercises animated in house
0AI calls or cost per generated workout
1Subscription covering two people

Like OneLab, Unio started as something we needed ourselves, and the answer was to build it rather than wait for someone else to. It is launching on iPhone, with Android to follow, and the engine underneath it is ours to keep improving.

Next project

Speedy Insurance

View →