Two students can ask the same question and get genuinely different explanations. Here's the simple idea behind it.
On signup you pick your age, goal, level and archetype. That profile travels with every question you ask.
A lightweight first pass separates real study questions from small talk, and pulls out the clean academic question.
If a student with your profile already asked something similar, you get a great answer instantly — at near-zero cost.
On a fresh question, Astra writes an explanation shaped by your age, goal, level and tone — then saves it for learners like you.
A 14-year-old starting out and a college student prepping for placements shouldn't get the same paragraph. By caching answers per student segment instead of globally, Classmate AI stays both personal and affordable — personalization doesn't make answers un-cacheable, it just caches them per profile.
Because common questions are served from cache, the expensive part — generating a brand-new explanation — only happens when it's genuinely needed. That discipline is what lets us offer Classmate AI free to start.