What does “Memory (Agent Memory)” mean?
Stored context an AI agent uses across steps or sessions to improve continuity.
How Memory (Agent Memory) Works
Short-term: the current conversation and task steps
Long-term: saved notes, facts, past results
Retrieval: fetch relevant memory when needed.
Practical Example
A personal assistant agent saves that you prefer morning meetings and window seats. Next week it retrieves those notes when scheduling or booking travel, so you do not repeat yourself.
Why It Matters
Without memory, an agent starts from scratch each time. Memory also raises privacy questions about what is stored and for how long.
Common Questions About Memory (Agent Memory)
Q: Is memory built into the model?
Usually not. It is stored outside the model, such as in a database, and fetched back into the prompt.
Q: Can it be wrong?
Yes. Stored notes can be outdated or mistaken, so good systems let users review and delete them.
Q: Is it the same as a context window?
The context window is the model's short-term working space. Agent memory persists beyond a single session.