What Makes This Different

I should be direct about this, because the landscape is full of apps making big claims.

Most language apps fall into a few categories. Gamified multiple choice (you know who). Flashcards with a skin. And increasingly: a ChatGPT wrapper with a conversation interface. That last category is growing fast, and it's worth being honest about what it is.

An LLM conversation partner can be useful. But it has no idea what you knew yesterday. It doesn't track which words you can recognise but can't produce. It doesn't know your double consonants are 40ms too short. It can't tell you why you freeze mid-sentence — whether it's a vocabulary gap, a grammar uncertainty, or a pronunciation fear. It just generates the next plausible response.

What's underneath Sentiero is different, and I think the difference matters:

A knowledge model that distinguishes recognition from production. Tapping a word for its meaning creates a debt. That debt is only repaid when you produce the word correctly, unassisted, in a later exercise or conversation. Until then, the word stays highlighted — a quiet reminder that understanding it is not the same as owning it.

Acoustic analysis, not a score. Thirteen diagnostic modules that measure specific, physical properties of your speech — duration, formant position, stress placement, pitch contour — and return findings you can act on. Not "7 out of 10."

A brain that explains itself. Every exercise has a reason, and you can see it: "We're practicing this because you stumbled on it in the café scene." Not because it's next in the syllabus.

Specificity to the crossing. This is built for Norwegian and English speakers learning Italian. The false friends, the sounds your ear skips, the grammar your mother tongue fights — every comparison and example is calibrated for that specific journey.

None of this is flashy. It's just the machinery that makes the difference between an app you use for two weeks and one that actually changes how you speak.