AI Scans · Capture
Unmonitored — Scans flowed into the model with no visibility, so nobody could tell how many ran or how well they performed.
GR1P — AI Fit Platform
Exarth built the web control center behind GR1P — an AI fit platform that goes from scan to perfect match. The dashboard monitors AI scans and the matching algorithm, and manages the product catalog, users, subscriptions, notifications and APIs.
- Scan & algorithm monitoring
- Tunable matching engine
- Product catalog & users
- Subscriptions & push alerts
- AI / ML
- UI/UX + Dev
- Admin
DashboardScansProductsSubscriptions
admin.gr1p.ai
The control center powering an AI fit experience — scans, matching, products and subscriptions.
AIScan → match
490+Scans monitored
296Products
LiveAlgo monitor
Overview/ 01
The control center for an AI fit platform
GR1P takes users from scan to perfect match. This web platform is its control center — monitoring AI scans and the matching algorithm, and managing products, users, subscriptions, notifications and APIs behind the experience.
Exarth designed and built the platform end to end — from the algorithm monitor and configuration to product, user and subscription management, RevenueCat billing and developer APIs powering the consumer app.
- AI scan & algorithm monitor
- Tunable matching configuration
- Products, users & subscriptions
- RevenueCat billing & push alerts
- Role
- UI/UX + Development
- Industry
- AI · SaaS
- Timeline
- 6 months
- Platform
- Web + Mobile
- Billing
- RevenueCat
- Year
- 2025
Scope & deliverables
Discovery & Strategy
Architecture for the scan-to-match flow, algorithm monitor and admin.
UX / UI Design
A focused dark design system for analytics, configuration and management.
Frontend Engineering
A fast, responsive dashboard with real-time scan and accuracy analytics.
Backend & Integrations
Algorithm config, product/user APIs, RevenueCat billing and push.
Challenge & solution
A black-box model, or a monitored, tunable engine
An AI matching model ran as a black box — no scan visibility, no way to tune it, and ad-hoc product, user and billing management.
- Unmonitored scans
- Black-box model
- Opaque billing
Algorithm · Matching
Black box — The matching algorithm was a black box — there was no way to monitor it or adjust its behaviour without a redeploy.
Products · Catalog
Static list — The product catalog the model matched against was static and hard to maintain, so matches went stale.
Users · Accounts
No admin view — There was no admin view of users, so support and analytics meant digging through the database by hand.
No core
No shared platform
Fragmented systems
0VisibilityAd-hocAdmin
Subscriptions · Billing
Opaque billing — Subscription state was opaque, so it was hard to see who was active, trialing or churned across stores.
Notifications · Engagement
No outreach — There was no way to reach users with timely alerts, so re-engagement relied on app-store updates alone.
Insights · Analytics
No reporting — With no reporting, scan volume, accuracy and growth trends were invisible to the team.
REST API · Developer
Closed — Integrations were ad-hoc with no documented surface, slowing the mobile app and partner work.
Measurable impact on the live platform
490+Scans
296Products
AIMatching
LiveMonitor
Key features
Everything behind the scan, in one console
From watching the algorithm in real time to managing products, users and subscriptions — every lever behind the AI fit experience lives in one console.
01/05 features
Auto-playing — hover to pause
admin.gr1p.ai/scans-algo-monitor Live
Feature 01Scans & Algo MonitorA real-time monitor surfaces total scans, valid/invalid counts and accuracy so the team can watch the model perform.
Engineering
Technology & Development
Engineered to power an AI fit experience — a monitored matching engine, product and user management, RevenueCat billing, push and a documented API.
System architecture
Mobile → API → matching engine → admin & billing
APP
Mobile AppScan capture · the consumer scan-to-match experience
iOS · AndroidAI
Matching EngineAI model that matches a scan to the best-fit products
Monitored · tunableUI
Admin ConsoleAlgo monitor · products · users · subscriptions · push
SSR · responsiveAPI
Data & BillingEncrypted data · RevenueCat billing · documented REST API
REST · RevenueCatHow we built it
6-phase delivery
- 01
Discovery & Architecture
Phase 1Mapped the scan-to-match flow, monitoring needs and admin into one blueprint.
- 02
UX / UI Design
Phase 2Built a focused dark design system for analytics, configuration and management.
- 03
Frontend Engineering
Phase 3Developed a fast, responsive console with real-time scan and accuracy analytics.
- 04
Engine & Integrations
Phase 4Engineered the matching config, product/user APIs, RevenueCat billing and push.
- 05
QA & Security
Phase 5Hardened auth and roles, then validated the monitor and matching end to end.
- 06
Launch & Iterate
LiveShipped with monitoring and analytics, then iterated on real model performance.
Technology stack
16 technologiesFrontend4
- Reactv18
- Next.jsApp Router
- TypeScriptv5
- Tailwindv4
Backend4
- Node.jsv20
- REST APIDocumented
- MatchingAI/ML
- RBAC
Data & Cloud4
- PostgreSQL
- RedisCache
- AWS
- Docker
Integrations4
- RevenueCatBilling
- PushAlerts
- Mobile AppScan
- AnalyticsMonitor
Product Screens
Scan to match, one console
Glide through the live GR1P console — dashboard, scans, algorithm config, products, users, subscriptions, push and APIs.
Drag to explore01 / 12
Philosophy
Design Approach
A focused, dark design philosophy built for clarity — turning AI scans and dense analytics into glanceable, actionable insight.
ClarityInsightTrustResponsive
- Design pillars
- 04
- Accessibility
- AA
- Analytics
- Live
- Audience
- Operators & developers
- Platform
- Web + Mobile
- Design system
- Dark · 30+ tokens
- Accessibility
- WCAG AA
Every screen is shaped around the model and its data — clarity first, insight at a glance, and a confident path from scan to match.
- Pillar 01
Data-First Layouts
LiveAnalyticsGlanceable analytics that turn scan volume and accuracy into instant insight.
Operators read model health at a glance.
- Pillar 02
Controllable AI
TunableEngineConfiguration surfaces that make a complex matching engine easy to tune.
The team adjusts behaviour without a redeploy.
- Pillar 03
Focused Dark UI
AAWCAG contrastA high-contrast dark interface that keeps charts and data crisp.
Easy on the eyes for long monitoring sessions.
- Pillar 04
Secure by Design
RBACAccess controlRole-based access protecting scans, users and billing data.
Sensitive data protected at every step.
Tokens
Design System
A cohesive dark visual language that keeps every analytics, configuration and management screen sharp and consistent.
5Colors
6Type sizes
5Spacing
8Icons
4Breakpoints
Color palette5
Typography6
Heading
Space Grotesk 800 · 32px
Subheading
Inter 500 · 20px
Body text for product listings, collections, and guides.
Inter 400 · 16px
Caption labels
Inter 400 · 14px
Small helper text
Inter 400 · 12px
Bold heading
Space Grotesk 700 · 18px
Components4
Add to cart Wishlist
Organic In stockHeirloom
⏎
DefaultHoverActiveOff
Spacing scale5
XS4px
SM8px
MD16px
LG24px
XL32px
More work
Explore more projects
More of what our team designs and ships — from AI and retail platforms to government systems and healthcare SaaS.
