About MCP Server

The Oracle Autonomous AI Database MCP Server is a managed, multi-tenant server designed to provide secure, standardized access to database tools and features through the Model Context Protocol (MCP). It provides an MCP server for each Autonomous AI Database, enabling AI agents and client applications to interact seamlessly with custom and built-in Select AI Agent tools using MCP APIs.

MCP Server integrates with database identity, authorization, and governance frameworks, supporting enterprise-grade security, auditing, and compliance. Exclusively available for Oracle Autonomous AI Database (versions 26ai and 19c), it eliminates the need for you to manage your own MCP infrastructure.

Topics

Parent topic: Autonomous AI Database MCP Server

Benefits of MCP Server

MCP Server offers several advantages that enhance the integration, security, and management of AI-driven applications within Oracle Autonomous AI Database.

MCP Server provides a secure and efficient way to connect AI Agents supporting MCP to the Oracle Autonomous AI Database.The key benefits are:

  • Enables seamless integration of Autonomous AI Database with MCP-compatible clients such as Claude Desktop, Visual Studio Code with Cline, and OCI AI Agent, while providing secure, governed, and scoped access.
  • No customer-side MCP server infrastructure involved, reducing deployment effort.
  • Integrated with database roles, supports network access control lists (ACLs) and Private Endpoint (PE) configurations.

Parent topic: About MCP Server

Feature Highlights of MCP Server

Explore key feature highlights of the MCP Server in the table below:

These features enable you to access MCP tools through agentic AI applications that support MCP clients. You must register the MCP tools with the DBMS_CLOUD_AGENT_AI package before you can access them.

Features Descriptions

Enable MCP server

A Database Administrator or the ADMIN user can enable or disable the MCP server using OCI free-form tags. See Enable MCP Server for more details.

Custom Tool Registration with Select AI Agent

Database users register custom tools using the DBMS_CLOUD_AI_AGENT PL/SQL package. See Create Select AI Agent Tools for more details.

Authentication

The MCP server supports authentication using database credentials. See Configure MCP Server in AI Agent Application for more details.

Security Controls

Implements rate limiting, Access Control List (ACL) and Private Endpoint (PE) enforcement, and fine-grained secured data access controls using Virtual Private Database (VPD) policies. See Create and Register a VPD Policy for more details.

Parent topic: About MCP Server

Architecture of MCP Server

This provides an overview of the MCP server architecture and its key components.

The MCP server architecture is organized into three key layers: Unified Security Layer, Multi-tenant MCP Server, and Select AI Agent Framework. Together, these layers provide a secure, governed, per-database MCP server that is natively integrated with Autonomous AI Database.

Unified Security Layer

With Unified Security Layer, you can:
  • Enforce network ACLs and support for private endpoints
  • Support for database identities
  • Enforce database roles and Virtual Private Database (VPD) policies

Multi-tenant MCP Server

With the Multi-tenant MCP server, you can:

  • Securely use Select AI Agent custom tools without managing MCP server deployment or infrastructure
  • Log all interactions to database audit logs

Select AI Agent Framework

With the Select AI Agent framework, you can:

  • Utilize tools defined through the Select AI Agent
  • Register and update tools
  • Manage the lifecycle of the tools that get exposed as MCP tools

See Select AI Agent for more details.

Parent topic: About MCP Server