A Pokédex-inspired wildlife scanner that uses computer vision and GPT-4o to identify animals from a photo — delivering species info, ecological context, and fun facts instantly.
Growing up, the Pokédex was the ultimate fantasy — point a device at something and instantly know everything about it. Anilog is that idea made real, but for the actual natural world.
The project started as a weekend experiment to see how far GPT-4o's vision capabilities could go with wildlife identification. The answer: surprisingly far. So I built it into a proper deployed app.
"What if you could Shazam a bird?"
Upload any photo and GPT-4o's vision model analyses the animal — works across species from common pets to rare wildlife.
Returns scientific name, family, habitat, diet, and conservation status — structured data presented in a readable format.
Fast API response times with a clean loading state. Built to feel snappy even on mobile connections in the field.
Anilog was my first proper end-to-end deployment — from idea to a live URL. It forced me to think about product experience beyond the code: how does a result page feel when you're standing in a park? Is the loading state reassuring? Does the information hierarchy make sense on a small screen?
It also showed me how quickly multimodal AI can compress what used to be a complex ML pipeline into a single, well-prompted API call.