Skip to main contentSkip to footer
    Back to Case Studies
    Fighter Cut+ The Corner for coaches
    Mobile AppCombat SportsIn-House Product

    Fighter Cut

    Make weight. Keep your power. Two apps, one backend: an athlete app that plans a safe cut, and The Corner, where a coach sees the whole fight team at a glance.

    Project Preview

    Live website preview - fightercut.com

    750
    kcal max daily deficit the engine will plan
    168
    Branded supplements compared, 65 brands
    236
    Seed foods with full macro data
    0
    AI API calls in the product

    Five-phase cut plan

    Built backwards from weigh-in, with sanctioned weight-class tables for MMA, boxing, wrestling, judo and jiu-jitsu.

    An engine that refuses

    Weekly loss graded safe, caution or unsafe. Impossible plans print the shortfall instead of a number. No sauna or water-cut protocols, ever.

    3D personal body model

    A turnable body solved from tape measurements, with Navy, 3-site and 7-site body-fat methods.

    Supplement advisor & compare

    A rule engine over 168 branded products, ranked by cleanness, value, dose, testing or vegan fit. Adding a rule needs no code change.

    Regulation-accurate labels

    Nutrition Facts and Supplement Facts panels with 21 CFR 101.9(c) rounding and from-scratch EAN-13 and UPC-A barcodes.

    The Corner

    Roster, weigh-ins and weekly load for every fighter in a gym. Invite codes, staff roles, and food logs that never cross over.

    The problem

    Making weight is not a diet. It is a deadline. Every combat athlete knows the shape of a bad camp: a plan built on a number that was never realistic, a last week of dehydration to cover the gap, and a fighter who steps into the cage having left their power on the scale. The tools on the market make it worse, because a generic calorie app will happily prescribe a deficit no sports physician would sign off on.

    Fighter Cut is a Frenchy Digital product, founded by Chris Machetto, and it started from a simple rule: the software should refuse to plan a cut that hurts the athlete. On the other side of the mat, coaches were texting a dozen fighters every Sunday to ask how camp was going. The Corner exists so nobody has to be texted.

    The design constraints we set ourselves

    • Safety limits in the engine, not the UI: Every diet week capped at a 750 kcal deficit. Weekly loss graded safe at or under 1.5% of bodyweight, caution to 2%, unsafe above. Intake floors of 1,500 kcal for men and 1,200 for women.
    • Honesty over a number: If the target cannot be reached safely, the app prints the shortfall and says so. The water cut is sized, at 3% by default, and handed to a professional. The app will not walk anyone through a sauna.
    • No AI coach: A fighter's body is not a place for a model to improvise. Every recommendation is deterministic, testable and explainable.
    • Private by default: No account is required. Data stays on the device unless the athlete joins a gym, and even then food logs never leave.

    Where AI fits, and where it deliberately does not

    There is no AI in the product. The athlete app makes zero AI API calls, and the App Store listing says "No AI coach" on purpose. The supplement intelligence is a rule engine over structured data, so a new rule is a data change, not a code change. The comparator scores labels deterministically. What AI did shape is how the product was built: both repositories were developed with Claude Code against a written engineering contract, which is how a five-phase planning engine, a 3D body model, a labelled supplement catalog and a second coach app shipped to TestFlight in under four weeks.

    How we built it

    Two apps, one backend, two philosophies

    Fighter Cut: local-first

    React 19 and Vite wrapped in Capacitor 8 for iOS. State lives in Dexie on IndexedDB with Zustand persistence, so a fighter can log food, water, sessions and tape measurements with no signal in a gym basement. Supabase only enters the picture for accounts, clubs and Pro billing.

    The Corner: server-owned

    A roster is server-owned by definition, so The Corner is the opposite: every screen is a live read from Supabase. Same auth, same project, same tables. All SQL lives in the athlete repo, and the coach app consumes it.

    Engines you can test

    Every number the app shows comes out of a pure function in an engines directory: basal metabolic rate, total daily energy expenditure, macros, the cut plan, projections, goals, body composition, the supplement comparator, readiness and a fasting state machine. Each is unit-tested in isolation. The charts are hand-rolled SVG, and even the barcodes on the label panels are rendered from first principles rather than pulled from a dependency.

    What shipped

    Five-phase cut plan built backwards from weigh-in, with sanctioned class tables (MMA and boxing shared, UWW, IJF, IBJJF with gi)
    Food, macro, water and session logging over 236 seed foods and a nutrient library
    14-site tape measurement, Navy, 3-site and 7-site body fat, and a turnable 3D body model
    Supplement advisor and comparison across 168 branded products from 65 brands, with affiliate buy links across 8 networks
    Eight meal-plan templates scaled live to the athlete's calories and bodyweight, and 15 formulas
    FDA-accurate Nutrition Facts and Supplement Facts panels
    Help centre with 78 articles and a ticketing flow
    The Corner: roster, fighter detail, assigned work, invite codes, staff management, gym claiming
    Founding programs: 50 founding athletes free forever, first 5 founding gyms free on the Team plan

    Timeline

    1. 9 August 2026: First commit of the athlete app.
    2. 24 August 2026: The Corner started as a separate repo on the shared backend.
    3. Late August 2026: Both apps in App Store Connect, internal TestFlight live, external beta group submitted for review. corner.fightercut.com live on the web.
    4. 1 October 2026: Compiled-in launch instant when paid gating arms.

    Problems worth talking about

    A body model from tape

    The 3D body is solved from circumference measurements using Ramanujan's ellipse-perimeter approximation, rotated as a true elliptical projection, with three deliberate anatomical overrides for the head, shoulders versus chest, and forearm versus wrist.

    Labels that agree with themselves

    Statutory rounding rules mean a declared value and its percent daily value can disagree if you round naively. The label engine rounds the way the regulation does, so the two always match.

    Coaches landing in the wrong app

    Supabase silently substitutes its Site URL for unlisted redirect targets. Coaches were landing in the athlete app holding a valid session, with nothing reporting an error. Found, documented, fixed.

    Keeping the admin console out of the binary

    A build flag resolves the admin console to a stub at module-resolution time, so it cannot ship inside the App Store build even by accident.

    Where it stands

    Beta, reported honestly

    Fighter Cut and Fight Corner are on TestFlight with public release gated to October 2026. The numbers on this page are counts from the codebase: foods, products, brands, articles, safety thresholds. We do not publish downloads, active users, gyms or revenue for an app that has not launched, and this page will be updated when it has.

    Making weight is not a diet. It is a deadline.

    — Fighter Cut App Store listing

    What This Project Does Well

    Local-first: the athlete app works fully offline with no account, on Dexie/IndexedDB and Zustand
    Every calculation lives in a pure, unit-tested engine: BMR, TDEE, macros, cut plan, projection, body composition, readiness
    Hand-rolled SVG charts and a hand-rolled barcode renderer: no chart or component libraries in the bundle
    Consent by design: only weigh-ins, sessions and completed assignments reach a coach, never a food log
    Fully headless iOS pipeline: xcodebuild plus the App Store Connect API, no clicking through Xcode
    Metric everywhere internally, converted once at the display layer

    Technologies Used

    React 19, Vite, TypeScript, Dexie (IndexedDB), Zustand with persistence, Capacitor 8 for iOS, hand-rolled SVG charts, Supabase (Postgres, Auth, RLS) for accounts, clubs and billing, RevenueCat Web Billing, Stripe, Cloudflare Pages, TestFlight and App Store Connect API automation

    Ready to Build Something Similar?

    Let's discuss how we can help transform your business with a custom digital solution.

    Visit fightercut.com📅 Schedule Consultation