Student memory for tutors that adapt instead of restarting.
Track learner profile, weak topics, exam context, learning style, language comfort, and progress signals so every session starts from the student's actual state.
Weak topic: integration by parts is still difficult
ranked, sourced, and retrieval-ready
Learns better from worked examples before theory
ranked, sourced, and retrieval-ready
Preparing for Class 12 board and JEE-style questions
ranked, sourced, and retrieval-ready
Comfortable with Hinglish explanations
ranked, sourced, and retrieval-ready
Learning problem
Why this needs memory infrastructure.
The hard part is not remembering text. It is deciding what is durable, what is stale, who controls it, and what the next agent should trust.
Tutoring agents forget the student's weak topics between sessions.
A student's learning style matters more than generic lesson retrieval.
Exam goals, grade level, and timeline need to survive across chats.
Operators need to inspect what was stored without reading full private conversations.
What you get
Clear context for the next model call.
Weak topics
Store concepts the student struggles with and how severe the gap is.
Learning style
Examples-first, visual, step-by-step, theory-first, or language-specific preferences.
Exam context
Remember grade, exam date, target, and readiness signals for future sessions.
Language comfort
Support English, Hindi, Hinglish, or other languages based on extracted preference.
Tutor prompt
Return a compact teaching profile the tutor can use before answering.
Structured schema
Use a domain-specific overlay instead of dumping all memory into generic facts.
Integration
Use the SDK first. Add governance when needed.
The same API supports solo apps and production teams. Start with a stable user ID, then add source metadata, service writers, or Memory Passport when your product needs them.
from memoryos import Memory
mem = Memory(api_key="mem_live_xxx")
mem.add(
external_user_id="student_001",
messages=lesson_messages,
metadata={"course": "calculus"},
)
profile = mem.get_edtech_profile("student_001")
context = mem.get(
external_user_id="student_001",
query="how should I teach integration by parts now?",
)Build a tutor that remembers how each student learns.
Start with EdTech Schema in a test workspace, run real learner conversations, and verify extracted memory before launch.