Zovex
QR-based restaurant ordering and management platform with multi-tenant architecture.
Overview
Zovex is a QR-based restaurant ordering and management platform that lets guests order and pay from their phones while restaurants manage tables, orders, and analytics. Each restaurant gets a pre-hosted website on a custom subdomain (e.g., foodmaster.zovex.app) with customizable branding and live menu integration.
Problem
Restaurants relying on paper menus, manual order-taking, and static websites face slower service (8+ minutes from order to kitchen), frequent errors, and fragmented online presence. They also struggle with the cost and complexity of maintaining their own website infrastructure.
Key Issues
- Manual order-taking increases wait times (8+ minutes) and leads to errors during peak hours
- Paper menus and static PDFs are hard to update and provide no behavioral data
- No unified system to manage tables, in-house orders, online orders, and payments together
- High cost and technical complexity of hosting and maintaining a restaurant website
- Restaurant owners struggle to maintain a modern, up-to-date website connected to their live menu
- Lack of real-time visibility into order status and kitchen operations
Solution
A full-stack multi-tenant SaaS platform where guests scan a table QR to view the live menu, place orders, and pay online. Each restaurant gets a pre-hosted website on a custom subdomain with customizable branding and real-time menu synchronization.
Implementation Highlights
- Architected and deployed a multi-tenant SaaS platform with real-time order management using optimized database queries and WebSockets
- Built scalable subdomain-based architecture with tenant isolation, integrated payment orchestration (BYOK support), and ML-powered menu recommendations
- Established testing pipeline with Pytest for backend services and Playwright for E2E flows; set up GitHub Actions for automated CI/CD deployment
- Optimized concurrent request handling and implemented monitoring with Sentry for production error tracking
- Implemented real-time menu synchronization so changes appear instantly across QR menus and public websites
- Designed QR-based ordering flow that eliminates manual order-taking and reduces errors
- Created live order management dashboard with real-time updates across all connected devices
- Developed ML recommendation system using collaborative filtering and content-based signals from order history
Outcome
The platform streamlines in-restaurant ordering with 16x faster order-to-kitchen times (from 8 minutes to under 30 seconds), provides instant online presence through pre-hosted subdomains, and lays the groundwork for data-driven recommendations.
Key Results
- Latency optimized architecture: Built a low-latency real-time order pipeline by optimizing database access patterns and introducing WebSocket-based push delivery.
- Instant online presence with pre-hosted websites on custom subdomains (e.g., restaurantname.zovex.app)
- Eliminated website hosting and maintenance costs for restaurant owners
- Improved order accuracy by sending orders directly from guest devices to the kitchen dashboard
- Centralized management of locations, tables, menus, orders, payments, and public website in a unified interface
- Enabled restaurants to accept online payments via their own gateways while Zovex handles orchestration
- Real-time updates across all devices using WebSocket implementation
- Handling daily orders for multi-table restaurant environments in production
Challenges
Designing a platform that combines subdomain routing, real-time order management, table management, online ordering, website customization, BYOK payments, and ML-driven recommendations while achieving sub-30-second order times required careful performance optimization.
Technical Challenges
- Implementing dynamic subdomain routing with Next.js middleware while maintaining clean separation between restaurant websites
- Building multi-tenant architecture where each restaurant's website is isolated but shares infrastructure
- Optimizing database queries and implementing WebSocket for real-time updates to achieve 16x performance improvement
- Modeling tables, orders, menus, and websites so they stay in sync while serving both in-house and online guests
- Keeping order and payment states consistent and traceable when using external payment providers
- Ensuring low-latency, real-time updates between guest devices, owner dashboards, and kitchen views
- Planning and implementing recommendation features while dealing with early cold-start and evolving data volume
- Building comprehensive testing pipeline covering unit tests (Pytest) and E2E flows (Playwright)
Key Learnings
The project provided hands-on experience in building a production-ready vertical SaaS for restaurants that combines subdomain routing, real-time order management, QR ordering, website building, payments, ML features, and comprehensive testing/deployment pipelines.
What I Learned
- Mastered performance optimization techniques reducing order processing time by 16x through database query optimization and WebSocket implementation
- Mastered dynamic subdomain routing using Next.js middleware for multi-tenant SaaS architecture
- Learned to design order and table lifecycles that work for both small single-location restaurants and more complex setups
- Deepened understanding of the cold-start problem and how to shape the UX and data model to collect high-value behavioral data for ML
- Gained experience implementing BYOK payment flows and reconciling external gateway states with internal order logic
- Refined patterns for building real-time, observable systems with proper error monitoring using Sentry
- Developed expertise in testing strategies with Pytest for backend and Playwright for E2E flows
- Mastered CI/CD pipeline setup with GitHub Actions for automated deployments