Oracle Brings Select AI to the Java Ecosystem with an Official, Production-Ready SDK
On September 22, 2026, Oracle quietly shipped one of the most consequential releases for Java-centric enterprise teams working with Oracle Autonomous AI Database: Select AI for Java 1.0.0. Published to Maven Central under the coordinates com.oracle.database.selectai:select-ai:1.0.0 and open-sourced on GitHub at oracle/java-select-ai, the SDK provides a clean, typed Java API that wraps the full breadth of Select AI capabilities—eliminating the need for developers to hand-code PL/SQL calls or build fragile JDBC wrappers around DBMS_CLOUD_AI.
For the millions of Java developers who build and maintain enterprise applications on Oracle databases, this release is a clear signal: Oracle wants AI-powered database features to be as natural to consume in Java as they are in SQL or Python.
What the SDK Covers
Select AI for Java 1.0.0 is not a thin wrapper or a proof-of-concept. It exposes a comprehensive surface area of Select AI functionality through typed Java classes and interfaces, organized around a central SelectAI client. The capabilities include:
- Natural Language to SQL (NL2SQL): Generate, execute, and narrate SQL queries from plain-English prompts—the flagship Select AI feature that lets business users and developers query data without writing SQL by hand.
- Retrieval-Augmented Generation (RAG): Integrate retrieval-augmented generation workflows that combine vector search with large language model responses, grounding AI answers in your actual enterprise data.
- Conversations: Manage multi-turn conversational interactions with the AI, maintaining context across exchanges for more sophisticated question-and-answer workflows.
- Vector Index Management: Create, configure, and manage vector indexes directly from Java code, supporting the semantic search infrastructure that underpins RAG and similarity-based retrieval.
- Summarization and Translation: Leverage built-in AI actions for summarizing documents or result sets, and translating content between languages.
- Synthetic Data Generation: Generate realistic synthetic datasets for testing, development, and compliance scenarios—all driven through the same unified API.
- Feedback: Submit feedback on AI-generated results to improve model accuracy and relevance over time.
- Profile and Credential Management: Programmatically configure AI profiles and cloud credentials, making it straightforward to automate deployment and environment setup.
Flexible Connection Management and Broad Compatibility
Oracle designed the SDK with enterprise deployment realities in mind. It supports two connection models: SDK-managed JDBC connections, where the library handles connection lifecycle internally, and application-managed DataSource connections, which let teams plug the SDK into existing connection pools, Spring Boot configurations, or Jakarta EE containers without friction.
The SDK requires JDK 17 or higher—aligning with the modern Java baseline that most enterprise teams have adopted or are actively migrating to. Crucially, Oracle has validated Select AI for Java against both Autonomous AI Database 26ai and 19c, ensuring that organizations still running long-term-support database versions are not left behind.
Completing the Language Parity Story
This release directly complements the Select AI for Python SDK, which Oracle released in September 2025. Together, the two SDKs ensure that the two dominant languages in enterprise Oracle environments—Java and Python—have first-class, officially supported paths to every Select AI capability.
The significance of this parity cannot be overstated. While Python has become the lingua franca of data science and AI experimentation, Java remains the backbone of production enterprise systems, microservices architectures, and mission-critical applications in Oracle shops worldwide. By offering a native Java SDK, Oracle removes a barrier that might otherwise force teams to introduce Python services solely to access AI database features, or to maintain brittle custom integration layers.
A Broader Push: Sharing AI Agent Teams Across Schemas
The Java SDK release arrives alongside another notable enhancement: the ability to share Select AI Agent Framework teams across database schemas. This capability expands multi-tenant AI workflow reuse, allowing organizations to define AI agent teams in one schema and make them available to others—streamlining governance, reducing duplication, and enabling centralized AI capability management in complex enterprise environments.
Taken together, these releases paint a picture of Oracle systematically removing friction from AI adoption at every layer: from the database engine, through the agent framework, and now into the application SDKs that developers use daily.
Practical Takeaway for Oracle Professionals
If your team builds Java applications on Oracle Autonomous AI Database, Select AI for Java 1.0.0 should be on your evaluation list immediately. Adding it to your project is as simple as including a single Maven dependency. Start by experimenting with NL2SQL generation and RAG workflows in a development environment—these are the capabilities most likely to deliver immediate, visible value to both technical teams and business stakeholders.
For architects, the dual-connection-model support means you can integrate the SDK into existing Spring Boot or Jakarta EE applications without rearchitecting your data access layer. And for teams already using the Python SDK, adopting the Java counterpart creates a consistent, officially supported AI integration strategy across your entire application portfolio.
Oracle is making it clear: AI-powered database capabilities are no longer features you access only through SQL worksheets or PL/SQL scripts. They are becoming native citizens in every language your enterprise runs on.