Long-Term Memory for Agents

Agents now build long-term memory of your users, so answers get sharper the more people interact with them. One centralized memory layer runs across every agent, holding onto the details that matter like how you refer to things, what you care about, and how you like answers framed. Users stay in control of what gets remembered, with the ability to review and remove stored details so memory stays accurate over time.

Read the full product documentation here.