PokeStarter

Inspiration

We were inspired by the Pokémon card craze and how overwhelming it can be for beginners. With so many packs, cards, and prices, we wanted to create a tool that makes getting started easier and more fun.

What it does

PokeStarter is an AI-powered assistant that helps new collectors figure out what Pokémon cards or packs to buy based on their interests and budget. Users input a natural language question and a budget, and the app responds with tips and real-time card suggestions using AI and pricing data.

How we built it

We built a Node.js + Express backend that connects to the OpenAI API and the Pokémon TCG API. The frontend was built using vanilla HTML, CSS, and JavaScript for simplicity and accessibility. The user inputs are sent to the backend, which fetches relevant card data and combines it with an AI-generated response.

Challenges we ran into

Learning the new structure of OpenAI’s v4 API. Getting native fetch working properly in Node.js. Parsing meaningful keywords from user questions to query card data accurately. Handling API errors gracefully between frontend and backend

Accomplishments that we're proud of

Successfully combining generative AI with real-time pricing data. Creating a friendly and informative experience for beginners. Clean integration between a custom backend and frontend. Building my first working full-stack app in a short amount of time

What we learned

How to structure full-stack apps with separate frontend/backend folders. OpenAI’s new v4 SDK and chat API usage. How to use .env securely with .gitignore. Working with third-party APIs like Pokémon TCG’s REST API

What's next for PokeStarter

Improve the key word detection possibly with AI. Continue to improve the frontend because it is kind of ugly. Fix current bugs.

Built With

• css
• html
• javascript
• node.js
• openai
• pokemontcg