Brawn

Brawn is a UK company that sells strength training to gym operators rather than to consumers. Its model puts four to six members with a personal trainer for a structured block of sessions, so a gym can offer coaching to members who would not book it one to one. The company describes itself as “the all-in-one growth engine for your gym”, one that “finds the members you're missing, keeps the ones you have, and turns both into PT revenue”. It says it is trusted by more than 50 gyms across the UK, among them The Gym Group, UltraFlex, 6Fit and Bad Company.

Members book and track their sessions in the Brawn app, which sets out to help people “Lift Better, Lift Together” through booking, digital strength challenges and progress tracking with suggested weights. Underneath that sits the part a gym never sees: subscriptions, credits, session scheduling, trainer earnings and payouts, all of it running across several gym brands at once.

Bitcube has worked as part of Brawn's engineering team on that platform since July 2025.

The platform was five and a half years old by the time this work started, and it was already taking money from paying members, so there was no quiet period in which to change it. Billing state lived in two places at once, in Brawn's own tables and in Maxio, the subscription service it charged through, and the two had to agree on ten different membership statuses resolved by priority and then by date. When they disagreed, a member was charged for sessions they could not book or booked sessions nobody was charging for, and nothing in the system said so out loud. Trainers were paid out of the same numbers, so a billing error became a wrong payout to a self-employed person who noticed. Each gym operator carried its own branding, pricing and payment account, which meant no fix could assume one configuration. And four separate applications ran against one API — the member app, the public booking site, an onboarding app for trainers and the internal admin panel — so a change to the booking rules had to hold in all four at once.

Reporting ran alongside all of it: PT earnings by brand and month with CSV export, payout summaries, then per-credit revenue, a monthly profit and loss view, and finally a revenue recognition schema reconciled against Maxio and refreshed live on its webhooks. Membership administration was built out in the same stretch — upgrades, reactivation, and moving a member between a block and a recurring plan, each with its credit and billing consequences handled by the system. In June 2026 dependency vulnerabilities were cut from 52 to 2, with no critical or high remaining.