Skills, not scores.

Most learning apps grade a subject as a single percentage. Sparqle doesn't. Underneath, every subject is a graph of small, teachable micro-topics wired together by prerequisite — you can't be recommended "subject-verb agreement" until the graph can see you've actually mastered the sentence-parsing and noun-paradigm skills it depends on.

That graph structure is what lets Sparqle's "what's next" recommendation mean something: it's not a random worksheet, it's the next unlocked node in the map of what your child already knows.

The taxonomy builds on the open-source Marble Skill Taxonomy. Czech language arts (Český jazyk a literatura) is modeled separately and directly against the official RVP ZV framework, rather than as a translation of the English content — the two curricula don't share a structure, so we didn't force one.