Oracle AI Database Integrations
Use this overview to choose an Oracle AI Database integration for retrieval, agent state, and AI application persistence.
Oracle AI Database integrations connect AI application frameworks to database capabilities such as vector search, document loading, checkpointing, and durable memory.
Available Integrations
Use the available integrations to match framework features with the database capabilities that your application needs.
| Integration | Use For | Primary Oracle Capabilities |
|---|---|---|
| LangChain | Retrieval-augmented generation, semantic search, chat history, semantic caching, and agent memory. | Oracle AI Vector Search, Oracle Text, in-database embeddings, vector stores, and message history. |
| LangGraph | Durable agent state, checkpoint persistence, long-term memory, and agent workflows. | Checkpoint tables, JSON stores, connection pooling, and vector search. |
| Spring AI | Spring Boot applications that need vector stores, similarity search, metadata filtering, and chat memory. | Oracle AI Vector Search, JDBC connectivity, vector indexes, document metadata filters, and chat memory persistence. |
Choose an Integration
Use LangChain when an application needs a framework for retrieval and model orchestration. LangChain works well when the main database tasks are document ingestion, chunking, embedding, vector search, retrieval, chat history, or semantic caching.
Use LangGraph when an application needs durable state across agent steps. LangGraph works well when the application must pause, resume, checkpoint, or search long-term memory across graph executions.
Use Spring AI when a Spring Boot application needs to store document embeddings, run similarity search through a Spring-native VectorStore abstraction, or persist chat memory with JDBC.
Use LangChain and LangGraph together when a stateful agent needs retrieval and durable workflow state. For example, an agent can retrieve context through LangChain components and persist graph state through LangGraph components.
Shared Oracle AI Database Capabilities
Oracle AI Database provides shared capabilities that support both integration patterns:
- Vector search: Store embedding vectors with source data and query them with semantic similarity search.
- Document processing: Load and process files with Oracle Text when a retrieval workflow needs source documents.
- Hybrid retrieval: Combine semantic and keyword signals when an application needs more precise retrieval.
- Durable persistence: Store application state, messages, and memory in database tables with transactional guarantees.
- Operational control: Use existing database security, backup, monitoring, and lifecycle processes for AI application data.
Related Guides
Use the related guides for implementation details, setup steps, and examples.
| Guide | Scope |
|---|---|
| LangChain and Oracle AI Database Integration Guide | Build retrieval-augmented generation, agentic search, semantic caching, chat history, and memory-augmented AI applications. |
| LangGraph and Oracle AI Database Integration Guide | Persist checkpoints, stores, agent state, and searchable long-term memory for LangGraph applications. |
| Spring AI and Oracle AI Database Integration Guide | Configure Spring AI Oracle vector stores and chat memory for Spring Boot applications. |