agent-memory

To add long-term memory to a LangGraph agent, you can use Databricks Lakebase to persist user facts and conversation history across multiple sessions. This requires configuring Lakebase resources and specific Python dependencies.

The best way to maintain conversation history within a LangGraph session is using AsyncCheckpointSaver for short-term memory. This retains context throughout the active interaction without needing external storage.

Yes, you need the databricks-langchain dependency to integrate memory capabilities into your LangGraph agents. You must also configure specific Lakebase resources and environment variables for full memory management.

An agent remembers user preferences across sessions by leveraging Databricks Lakebase for long-term memory. This allows the agent to recall previously stored user facts like preferred names, teams, or past requests.

Yes, you can use AsyncCheckpointSaver with Databricks Lakebase to provide both short-term and long-term memory. AsyncCheckpointSaver handles within-session history while Lakebase handles cross-session persistence.

Your LangGraph agent forgets user facts between sessions because it lacks long-term memory configuration. You must integrate Databricks Lakebase to persist user facts and conversation history across sessions permanently.