What is intuitive-tarot-production?

This is a web application that offers AI-powered tarot card readings. A user types in a personal question, draws tarot cards from a virtual deck, sees the actual Rider Waite Smith tarot card images, and receives an AI-generated interpretation that takes both the cards drawn and the question into account. The project describes the experience as reflective rather than predictive, a structured way to explore emotions and choices rather than a tool for making absolute predictions. The app is split into two parts deployed separately. The frontend is a static website hosted on Vercel, built with HTML, CSS, and JavaScript. The backend is a Node.js server hosted on Render, and its main job is to act as a proxy between the browser and the AI model provider. This separation means the AI provider's API key is never exposed to the browser, the browser only talks to the project's own backend, which holds the key securely in server-side environment variables. The public demo version includes additional protections to limit API costs: an access-code modal that users must complete before getting a reading, and per-IP daily rate limiting that caps the number of readings each visitor can request. Automated tests check that these protections are in place, that the frontend does not accidentally expose API keys, and that the backend enforces the configured model instead of trusting whatever the browser sends. The tech stack is HTML, CSS, JavaScript, and Node.js, with Vercel for the frontend and Render for the backend. The project is intended as a complete, portfolio-ready AI web application rather than a mockup, covering both user experience and deployment considerations.