AI agent glossary
Agent memory
Any mechanism that lets an agent keep useful information beyond a single conversation, such as notes files, databases, or retrieved past sessions.
Language models do not remember anything between calls. Everything an agent knows in the moment has to be in its context window. Agent memory is the set of techniques for deciding what to save, where to save it, and how to bring it back when it matters.
It helps to separate short-term memory, which is the current conversation, from long-term memory, which persists across sessions. Long-term memory can be as simple as a markdown file the agent reads at startup and appends to, or as elaborate as a database the agent queries with a search tool.
The hard part is not storage but curation. Memory that grows without pruning fills up with stale or wrong facts, and the agent will trust them. Good memory systems save small, specific facts, record where they came from, and let the agent or a person correct them.
A personal ops agent keeps a preferences file noting that you prefer meetings after 10am and never on Fridays. Each morning it reads that file before proposing your schedule.
Learn more in the course
Related terms
Go from definitions to doing
Learn to use and build AI agents in 65 practical lessons.