Service Booking Assistant — Full-Stack AI Application

I built a full-stack booking application demo for a service-business workflow. It brings service discovery, booking handling, AI-assisted responses, and staff operations into one experience.

Service businesses need a clear path from exploring services to managing booking requests. They also need an operations view for tracking bookings and routing cases that require human follow-up.

Visitors can browse the service catalog, check appointment availability, review a booking summary, confirm a request, retrieve it by reference, and cancel it. Cancellation releases the appointment slot for a later request.

Google Gemini receives catalog and availability information from the backend and supports booking-preview and human-handoff actions. Gemini requires credentials and configuration.

The staff console provides booking lists, status changes, internal notes, and a human-handoff queue. The administrator area uses API-key-based access for the demo.

I handled the full-stack implementation: responsive interface work, backend API development, persistent-data setup, booking workflow design, Gemini integration, and the staff-facing operations interface.

Next.js, React, TypeScript, Tailwind CSS, Python, FastAPI, SQLAlchemy, Google Gemini, and SQLite. PostgreSQL and asyncpg configuration are included; PostgreSQL execution and concurrent-booking behavior have not been verified.

The portfolio includes screenshots of the service catalog and the staff console.

Independent portfolio demo for a fictional service business, using demonstration data. Not presented as production-ready.

The fallback without Gemini does not persist a real ticket. API-key administrator access is for demonstration purposes and is not production authentication. The AI action-confirmation flow needs further hardening before production use.