Resto Master
A multi-tenant restaurant platform for reservations, menu, orders, customers, team roles and an AI assistant, with Stripe subscription billing.
- ROLE
- Founder, product designer and builder
- TECHNOLOGY
- React 18TypeScriptViteFirebase AuthFirestoreStripeAnthropic APIVercel Functions
PROBLEM
Restaurants run reservations, menus, orders and staff across disconnected tools. Friction lands on the team during service and on the guest.
OPPORTUNITY
One connected system can cut that friction, with role-based access so each person sees only what their job needs.
CONCEPT
Tenant-isolated modules for Dashboard, Reservations, Menu, Orders, Customers and Settings, plus an AI assistant that runs server-side. Modules still on sample data are labelled “Preview module” in the app.
RESEARCH
Shaped by hospitality and operations experience. Structured interviews with working operators are a next step.
DESIGN
An operator-first interface, with roles that are enforced in the database rules and not only in the UI. Admins invite managers and employees, and nobody can choose their own role.
DEVELOPMENT
React 18, TypeScript, Vite and Tailwind with shadcn/ui on the frontend. Firebase Authentication and Cloud Firestore hold the data, protected by security rules. Vercel Functions handle AI chat and Stripe checkout, portal and webhooks. Testing covers unit tests, Firestore rules on the emulator, Playwright end-to-end with accessibility checks, and a Content Security Policy test against the production headers.
CHALLENGES
- Scope: restaurant operations are broad, so the first live modules were chosen carefully.
- Access control: roles had to hold in the security rules, not just the interface.
- Keeping the AI assistant server-side, so API keys never reach the browser.
ITERATIONS
- Core modules on real tenant-isolated data
- Team invites and roles, enforced by Firestore rules
- Billing through Stripe Checkout and the Billing Portal
- Preview modules for kitchen, POS, inventory and more
CURRENT STATUS
Live modules: Dashboard, Reservations, Menu, Orders, Customers, Settings and the AI assistant. Kitchen, Bar, POS, Analytics, Staff, Inventory, Loyalty, CRM and Marketing are preview modules with sample data.
LESSONS
- Put access rules where they can't be bypassed, in the database and server functions.
- Label preview features clearly so nobody mistakes sample data for real reporting.
NEXT STEPS
- Replace the preview modules with real data, starting with Kitchen and POS
- Pilot with working restaurants and gather operator feedback
- Move remaining sample-data screens onto tenant-isolated storage