All projectsCase study 02 / 07

/ AI SaaS Platform | 2026

Zentiva

A credit-based AI chat product combining persistent conversations, authentication, usage control, and Stripe Checkout.

Zentiva project interface

/ Project overview

Zentiva is a full-stack AI SaaS experience built around an authenticated chat workspace. Users can create conversations, revisit saved threads, interact with Google Gemini, track their remaining credits, and enter a Stripe Checkout flow from the product's pricing experience.

/ Process

From challenge to working product.

/ Phase01

The challenge

An AI chat interface becomes a real product only when identity, persistence, usage limits, billing entry points, and application state work together. The challenge was designing those systems as one coherent experience rather than treating the model call as the entire application.

/ Phase02

The approach

Clerk handles authentication while Prisma and PostgreSQL store users, threads, messages, and credit state. The backend checks available credits before calling Gemini, stores completed exchanges, and deducts usage after successful responses. Stripe Checkout Sessions provide the payment-entry workflow.

/ The result

A focused foundation designed to scale with the product.

The project demonstrates the product infrastructure surrounding an AI model: persistent user data, authenticated access, credit-based usage rules, payment integration, and a polished SaaS workspace.

/ Next project

Northline