Configure your agent’s memory settings in the Workbench to control how it stores and recalls information.
Agent Memory Configuration

Memory Types

Each agent includes a PostgreSQL database for persistent storage. Configure which memory types your agent supports using the toggles in the Workbench UI:

Session Storage

Automatic conversation history and state persistence (always enabled)

User Memories

Long-term facts and preferences about users (“ChatGPT-style” memory)

Session Summaries

Automatically summarizes long conversations for future reference
For advanced use cases, the storage is accessible programmatically via agno_agent.storage.

Using Memory in Your Code

The memory configuration from the Workbench UI is automatically applied to your agent. Use the user_id parameter to access user-specific memories and conversation history:
from xpander_sdk import Backend
from agno.agent import Agent

backend = Backend()
agno_agent = Agent(**backend.get_args())  # Memory settings applied from UI
agno_agent.print_response(message="What did we talk about yesterday?", user_id="user123")