I’m an indie dev and built VeriIA (https://detectordeia.pro), an AI detector for Spanish and English text.

The motivation: there are plenty of AI detectors for English, but I couldn’t find much for Spanish or other languages. Most people I know just paste Spanish text into an English-focused tool and hope the result still means something. I wanted to try what it looks like if you start from Spanish instead.

Right now the app is simple: you paste or upload a text, it gives a probability that it was written with AI, plus sentence-level highlights so you can see which parts look more “AI-like”. It’s mainly being used on essays and reports. The intention is to provide a rough signal for reviewers, not a definitive answer.

On the tech side it’s a fairly standard web stack: Next.js, React and Vercel on the frontend, with Supabase for auth and storing user scan history.

This is an early version: only Spanish and English are supported, longer texts work better than short ones, and the scores are obviously probabilistic and sometimes wrong.

I’m interested in feedback on two things: how (or whether) you think AI detection should be used in real workflows, and how you would approach evaluating a tool like this for non-English languages.