About Verb Engine

I built Verb Engine because learning Spanish conjugation from tables never clicked for me. Tables tell you what to say — they don’t explain why. Once I started seeing the endings as stacked signals (person, theme vowel, tense), conjugation went from memorisation to reasoning.

This app is the tool I wish I’d had: pattern-first, deliberately opinionated — short pages, no fluff, one clear mental model per surface.

The raw data comes from Kaikki.org, a structured extraction of Wiktionary — roughly 119k Spanish lemmas and 556k conjugated forms. Most of it arrives messy: phrasal idioms counted as separate verbs, archaic forms, regional tags sprinkled everywhere. Claude and I cleaned and reshaped it into a proper conjugation schema, then loaded it into Postgres on Supabase. The app itself is a small Next.js frontend on top.

Honest caveat: I’m iterating in the open, so expect rough edges. Feedback below genuinely shapes what I build next.