Spotter
Full-Stack Developer
About the Project
Spotter (internally Squotch) is a multi-tenant SaaS for independent gyms — a digital home for members and an operations console for staff. It began as a single-tenant build for Fit-Farm, a strength-and-conditioning facility on East West Road in Port Harcourt, and was generalized into a platform that onboards additional gyms while keeping Fit-Farm as the anchor tenant. It runs as three applications: a member-facing client, a staff admin console, and a NestJS API.
What It's Made Of
Member app
Next.js 16 · React 19 · Tailwind v4 · PWA
The gym member's surface, installable as a PWA: class schedules and bookings, check-in, membership and subscription status, payment history, and personal attendance history. Built as a PWA because members use it at the door on a phone, often on patchy data.
Staff & owner console
Next.js 16 · React 19 · Tailwind v4
The operations portal: member management, class and trainer scheduling, attendance oversight, subscription and payment administration, and the sales and revenue reporting an owner needs. Owner and staff are distinct permission levels here, not one 'admin' role.
API
NestJS 11 · Prisma 6 · PostgreSQL 16
The multi-tenant backend, where tenant isolation is enforced at the data layer rather than trusted to every query remembering a filter. It owns members, memberships, classes, bookings, attendance, payments and reporting, and it is the piece that was generalised from one gym to a platform while the anchor tenant stayed live on it.
Key Highlights
- Generalized a single-tenant product into multi-tenant SaaS without disrupting the anchor tenant's live operations
- Built member management, class scheduling, and attendance tracking
- Implemented role-based access control separating owner, staff, and member capabilities
- Shipped payments and subscription handling, plus sales and revenue reporting for gym owners
- Modelled the schema so tenant isolation is enforced at the data layer rather than in application code
Technical Challenges
The interesting work was the migration from single-tenant to multi-tenant. The original data model assumed one gym, so every table needed a tenant boundary added without breaking the anchor tenant already running on it. I chose to enforce isolation at the data layer rather than trusting every query to remember a WHERE clause — that decision cost more up front but meant a missed filter in application code could not leak another gym's members.