About Oracle AI Database

Learn how Oracle AI Database supports Oracle VecDB clients, vector workloads, and common AI application use cases.

Oracle AI Database is optimized for vector workloads and performs database management tasks automatically.

The Vector Database workload type offers Oracle AI Database capabilities in a configuration designed for AI-powered applications. You can use Oracle AI Database to manage vector data and build AI-powered applications.

Oracle VecDB supports access through the Python SDK, direct REST clients, and SQL or PL/SQL clients. Python SDK and REST client requests use the VecDB REST API through Oracle REST Data Services (ORDS), which invokes the DBMS_VECTOR_DATABASE PL/SQL package. SQL and PL/SQL clients can invoke the package directly.

The package creates, manages, and queries vector database objects, including vector tables, indexes, models, and vector data.

Description of the illustration vecdb-overview-workflow.png

Supported Workloads

You can use Oracle VecDB to:

• Store and retrieve vector embeddings efficiently for similarity search at scale.
• Build semantic search, RAG pipelines, and agentic applications that require durable memory that you can query.
• Combine vector similarity search with metadata filtering for more accurate, real-world results.
• Start quickly and scale as needed. You can upgrade to broader Oracle AI Database capabilities as workloads grow.

Common use cases include:

• Semantic search over documents, tickets, policies, and product catalogs.
• RAG for chatbots and knowledge assistants.
• Recommendations and similar-item experiences.
• Agent memory management for long-running agents across sessions and tasks, including episodic memory, semantic memory, and workflow memory.