Memory

What Memory Is

Memory is how the AI Agent retains information across separate conversation sessions. Unlike chat history — which is a record of the full message thread — memory stores specific facts or context that the agent can draw on in future conversations, without you needing to re-state them each time.

For example, if you tell the agent that your organization uses a specific naming convention for servers, or that a particular saved query represents your primary scope of interest, it can remember those details and apply them to future queries.

Memory is per-user and per-session scope. Each user's memory is independent and is not shared with other users. Admin users can turn off memory between chats so that content from one chat is not carried over to another.

What the Agent Remembers

The agent stores context you share that's likely to be useful in future sessions — things like environment details, preferences, or standing context about your work. It doesn't automatically remember everything, only what you provide explicitly or what it determines is worth retaining.

How Memory Differs from Chat History

Chat history is the full record of a conversation. Memory is a distilled set of facts that persist across sessions.

When you start a new conversation, the agent doesn't have access to previous chats, but it does have access to anything stored in memory. To reference something from a past conversation, describe it in your message or ask the agent to retrieve it.