Leverage the Oracle Autonomous AI Database MCP Server to seamlessly connect MCP-compatible clients with your database eliminating the need for customer-hosted MCP server infrastructure.
Oracle Autonomous AI Database MCP Server is a multi-tenant, built-in feature of Autonomous AI Database Serverless and Dedicated Region, providing MCP endpoints to your 26ai and 19c database instances.
In agentic development environments, the MCP Server lets us inspect the database, read from any schema, create new tables, generate sample data for development, and even write full migrations. Combined with existing back-end and front-end code generation, this is what makes truly agentic development possible.
Cofounder and CTO, Retraced
Honors core database security mechanisms, including RBAC, auditing, VPD, and redaction, when accessing database features through MCP endpoints
No additional infrastructure is required for the MCP server, and no MCP server installation is needed on a client machine
Benefit from Oracle’s automated patching and updates to ensure you have the latest features and security enhancements without any downtime
Manage and monitor MCP capabilities directly within the OCI console alongside your other database operations without needing separate management tools or experiences
Utilize Oracle's built-in compliance features to meet corporate and regulatory requirements, supporting auditing, data residency, and industry standards within the MCP Server's design
Integrated resource management and co-location ensure AI tasks and queries run with predictable performance, eliminating network latency and data movement common in external solutions
Easily create and manage custom or built-in AI tools via the Select AI Agent framework for a streamlined interface without the need for external backend or MCP infrastructure
Individual database users have a private set of tools through one fully customizable MCP server and sharable function implementations
Experience peace of mind with Oracle-managed services, featuring Oracle Premier Support and enterprise SLAs
Describe your desired application schema or data model in plain language to let your LLM generate the DDL and deploy it to your Oracle Autonomous AI Database using secure MCP server tools via Select AI Agent
Leverage the EXECUTE_SQL tool to access real-time performance and observability data, enabling your agent to analyze metrics, AWR/ASH data, and SQL execution patterns, while utilizing a RAG tool for comprehensive anomaly detection and targeted fixes
Build your application using custom tools to access and modify your database schema objects, and get started quickly with a wide range of tools supporting sample AI agents from our GitHub repository
Empower your back-office staff by transforming plain language questions into SQL queries on the Oracle Autonomous AI Database, simplifying tasks like ad hoc queries, long-term planning, forecasting, and pricing analysis with secure access to your business data through the built-in MCP Server
We are excited to announce the general availability of Oracle Autonomous AI Database MCP Server for Autonomous AI Database supporting database versions 19c and 26ai.
Oracle Autonomous AI Database MCP Server allows AI applications to seamlessly interact with the Autonomous AI Database through the MCP standard, removing the need for custom integration
Explore the Machine Learning blog for announcements, best practices, and tips on utilizing Oracle's in-database AI capabilities
In this Autonomous AI Database Learning Lounge session on the Oracle Autonomous AI Database MCP Server, PM and Development introduces the Autonomous AI Database MCP Server - a multi-tenant, built-in feature of Autonomous AI Database Serverless 26ai and 19c that exposes Model Context Protocol (MCP) endpoints. It enables AI agents and clients such as Claude Desktop, VS Code with Cline, and OCI AI Agent, to invoke tools you define with the Select AI Agent framework.
Discover how to set up your Autonomous AI Database as an MCP server and integrate it with Select AI Agent-defined tools







