Memory & Context: Long-Term Persistence

Explore how agents remember across sessions using vector databases and Firestore

4 lessons

ed232e48...

93a861b4...

content

Agent Memory: Beyond Conversation

Educational content slides

Agent Memory: Beyond Conversation

Chatbot memory (short-term): Conversation history within one session Close tab = forgot everything

Agent memory (long-term):

User preferences and context
Past tasks and outcomes
Knowledge gathered over time
Relationships between information Close tab = remembers you tomorrow

Example: Day 1: "I'm studying calculus for the AP exam" Day 7: "Find me practice problems" Agent: "Here are calculus practice problems focused on AP exam topics" ✓ Remembered: subject, purpose, difficulty level

1 / 6

54cfa73b...

quiz

Quiz: 3 Questions

Test your understanding with this quiz.

1 / 3

Why do agents need long-term memory beyond conversation history?

6be665e3...

feedback

Design Agent Memory

Complete this exercise and get AI-powered feedback.

Design Agent Memory

Create a memory system that balances utility (remembering useful context) with privacy (not storing sensitive data unnecessarily).

Consider: What makes interactions more helpful? What could be harmful to remember?

Design a memory schema for a 'career advisor' agent. What should it remember about users? Structure it into categories.

What should this agent explicitly NOT remember? How long should data be retained?

← previousMulti-Step Planning: Breaking Down Goals next →Build A Research Agent