ReadMyCup

AI-Powered Coffee Cup Reading & Tasseography Platform

Next.jsReactTypeScriptMongoDBOpenAI GPT-4oStripeNextAuthPostmark

The Problem Solved

Traditional coffee cup tasseography is an artisanal, manual practice unavailable to a global audience. The technical challenge was interpreting complex visual patterns in user-submitted photos and converting them into structured, coherent readings without prompt hallucination.

Architectural Scope & Implementation

Engineered an automated tasseography platform. Architected image processing pipelines utilizing GPT-4o vision to generate structured 5-part interpretations (Summary, Past, Present, Future, Advice). Built NextAuth session management, Stripe pay-per-reading credit bundles, and an administrative telemetry dashboard tracking token usage and Postmark delivery metrics.

Key Technical Challenge

Mitigating vision API latency and prompt drift, ensuring users receive consistent, structured JSON responses within tight request timeout limits.

Outcome & Engineering Result

Successfully launched with high reading satisfaction metrics, reliable credit processing, and real-time AI response delivery.

System Capabilities & Architecture Highlights

Computer vision pipeline converting coffee ground images into structured 5-tier readings

Pay-per-reading credit token architecture powered by Stripe checkout

Admin analytics panel tracking OpenAI token costs and email campaign engagement

Strict schema enforcement on LLM outputs to guarantee reliable rendering

Engineered by Zain Ali (HeyZain)

This system is part of Zain Ali's production portfolio. Zain is a full-stack developer specializing in Next.js, React, Node.js, TypeScript, and MongoDB, engineering full-stack platforms from schema design to edge deployment.

Read About Zain Ali's Background -> Explore Technical Writing & Research ->

Back to Projects Overview Back to Portfolio Home ->