Oracle updates AI Agent Memory Python library

Keeping agent memory inside the database is a smart operational play that prioritizes security, though it further locks developers into the Oracle ecosystem.

• –Custom extraction logic gives developers granular control over what facts are committed to long-term memory or kept in session context.
• –Hybrid search combining semantic vector search with keyword and metadata filtering dramatically improves retrieval precision.
• –Storing thread summaries and long-term memories in a unified database layer reduces the architectural complexity of mixing multiple databases.
• –While security and data governance are built-in, developer adoption remains gated by Oracle Database dependency.