Teaching that starts with understanding the learner
Seshvi combines a structured mathematical knowledge graph with AI reasoning, so every decision — what to ask, how to teach, whether a concept is mastered — is grounded in what a specific student has shown.
AI-powered learning diagnostics
An AI diagnostician tests self-reported confidence, adapts every question and stops when it has enough evidence.
Live today in Fractions
Misconception identification
Separates what a student did from why — e.g. adding numerators and denominators signals a missing idea about equal-sized parts.
Live today in Fractions
Individualized instruction
The AI tutor chooses explanations, analogies and visual models, and switches strategy when one does not work.
Live today in Fractions
Adaptive assessment
Mastery is judged from fresh problems and the student's own reasoning, not a fixed percentage.
Live today in Fractions
Personalized remediation
Pathways send students back to the prerequisite that is actually missing, and fast-track what is secure.
Live today in Fractions
Human-AI instructional collaboration
When the AI has tried several approaches without success, Seshvi suggests a human tutor.
Referral live; tutor matching coming soon
Available now
The full AI loop for Fractions (13 objectives, F01–F13), with parent accounts, multiple learners and cross-device progress.
Where it can go
The same knowledge-graph approach can extend to the other 32 topics in our curriculum, further grade levels and subjects beyond mathematics. These are future plans, not current features.