Pokédex
Using AI as a copilot to build a classic Pokédex web app.
I saw this YouTube video (link) where Gemini 2.5 Pro one-shotted a Pokédex web app.
Cool concept, right?
I personally loved it, but it looked nothing like the classic Pokédex I remember. So, I thought, “Let’s do it properly.”
I decided to set one ground rule.
1) Instead of having AI write the whole thing, I should leverage Gemini 2.5 pro as a genuine copilot.
So, could I collaborate with AI to build a Pokédex that would make a Pokémon master proud?
This is the result:
This front-end web app was built with vanilla HTML, CSS, and JavaScript, leveraging the PokéAPI for data.
The goal wasn’t just to replicate, but to see how the process of building changes when AI is a partner.
It’s less “AI builds a Pokédex” and more “Human and AI build a Pokédex”.
Want to try it?
Why not tap on the grass by my name in the tope left of this site?
The result is also hosted live at pokedex.maliglow.com and the code is fully open source over on GitHub if you want to embed on your own site.
But that’s not all
I wanted to test my hand at making a little promo trailer for the project, so I found rotato, and taught myself how to make a 3d render.
It turns out, getting the details right, like adding the actual Pokémon cries fetched from the API, makes a big difference.
It’s not a 1:1 replica, but definitely closer to the real deal than that initial demo. Plus, this one’s hosted reliably and predictably, partly thanks to lessons learned from my self-hosting adventures.