Table-Tap

Table-Tap is a high-performance, multi-tenant SaaS platform designed to modernize restaurant operations. It bridges four user environments under a single deployment:

• Customer Ordering (/order): Scanning table QR codes to browse menu categories, pay instantly via Paystack, and place orders.
• Kitchen Display (/kitchen): Real-time, touch-optimized order board for kitchen staff to coordinate and track orders.
• Restaurant Dashboard (/admin): Managing menu items, categories, viewing live analytics, and provisioning credentials.
• Super Admin Governance (/myadmin): Platform control panel (God Mode) to toggle paywalls, adjust pricing, and suspend tenants globally.

I designed and developed the fullstack architecture using Next.js and Supabase. Key contributions included:

• Secure Billing Loop: Connected client-side Paystack payment flows with Next.js Server Action verification. Validated transaction callbacks via HMAC SHA512 signatures and promoted order status in the database using a secure Supabase Admin Client.
• Two-Layer Security & AuthGuard: Created an edge-to-client guard system. Next.js Middleware verifies session cookies at the server edge, while a custom React AuthGuard performs deep, deferred checks for subscription validity and tenant suspension.
• Real-time Order Streams: Configured Supabase Realtime WS channels to stream PostgreSQL data changes directly to kitchen touchscreens, keeping order feeds perfectly synchronized.

• Decoupling Security for Performance: Running tenant verification queries in global Edge Middleware added heavy latency to page assets. I split the routing security layers so edge middleware only handles quick token refreshes, and the heavy database check runs inside the React AuthGuard shell.
• Preventing Transaction Spoofing: To prevent users from editing database order states manually, I forced the client to create 'PENDING' records, leaving status promotion to a secured webhook endpoint backed by cryptographic signature verification and database idempotency checks.
• Optimistic Credential Write: Listing all platform users to ensure uniqueness when provisioning device credentials created latency. I restructured the flow to write the record first, intercept duplicate key exceptions, and fallback to updating passwords, cutting database roundtrips to a single path.