Lathoa: Math learning app for kids where AI intentionally makes mistakes.

What it does

Lathoa is a math learning app where an AI character named Errol solves math problems step-by-step, intentionally making one mistake per problem. Kids identify the wrong step and write an explanation of what went wrong. The app categorizes errors into three difficulty tiers: Obvious, Sneaky, and Expert. Kids earn ranks (Rookie through Sherlock) based on catches rather than speed, though speed does contribute to points. The app shows all work step-by-step and tracks past problems as a "case file" that can be replayed and annotated.

Who it is for

The app targets children ages 10 to 14. It's positioned for parents seeking to develop critical thinking skills in their children, particularly the ability to spot errors and misinformation from AI systems. The interface is designed for homework-table use rather than algorithm-driven engagement.

Pricing

Free tier offers three cases per day with one kid profile and access only to Obvious difficulty. The Learner plan costs $66.99 per year for one child with full access. New accounts receive the Parent plan free for 7 days without requiring a card. A "Founding 50" offer provides 50% off the first year for the first 50 families.

How it stands out

The core mechanic inverts traditional learning app design by rewarding error detection over correct answers. Kids must explain their reasoning rather than simply selecting the right answer, which requires engagement beyond pattern recognition. The app avoids engagement tactics like leaderboards (unless parents enable them), badges for speed, or algorithmic personalization. It explicitly respects COPPA compliance, contains no ads, and does not harvest user data. The emotional feedback from Errol's reactive character provides intrinsic motivation without gamification pressure.

What a founder should check

A rival builder should investigate retention rates for error-detection gameplay versus traditional answer-focused math apps, since the mechanic is unfamiliar to users accustomed to immediate feedback for correct solutions. Examine switching costs: whether parents perceive sufficient differentiation from established math tutoring apps like Khan Academy or IXL to justify subscription costs and behavior change. Finally, verify the moat around AI mistake generation—whether calibrating difficulty tiers and preventing gaming (kids always guessing "there's an error") remains technically defensible as competition enters, or whether the novelty diminishes as competitors copy the format.

Thinking of building something like this?

Every launch here is a competitor to somebody's idea. If yours is close, check it against the market before you build: the Full Check names the rivals, the prices and the gaps.