Oracle Cloud Infrastructure Documentation

Generative AI

OCI Generative AI is Oracle's Enterprise AI Platform for building, deploying, and governing AI applications at scale. Access pretrained models or import your own, connect agents to enterprise data, and deploy applications on managed infrastructure. Combine tools, memory, retrieval, and isolated sandboxes to support your workflows, with enterprise controls for access, networking, and model behavior.

What's new

Get Started

Overview
Learn about using the models
Learn about Enterprise AI agents
Product page

Learn About Models

Pretrained models
Models by region
Hardware unit shapes by region  New
Routing profiles  New
Model discovery  New
Switch a hosted model  New
Import your own models
Guardrails
Retiring the models

Import Models

Compatible models
Import models from Hugging Face
Import models from Object Storage buckets

Tutorials

Route model requests across OCI regions  New
Compact conversation context  New
Create a sandbox user
Set up API authentication
Chat using LangChain

Build Agents with Enterprise AI features

Overview
Quick start guide for building agents
OCI OpenAI-Compatible Endpoints
OCI Responses API
Projects

Deploy agents with Enterprise AI features

Learn about deploying agents
Create an application
Create a deployment
Add an artifact

Developer Tools

Inference API
Management API
NL2SQL API
Inference CLI
Management CLI
NL2SQL CLI
SDKs and the CLI
Cloud Shell

Support

Support Requests
Create a service request

Oracle Cost Advisor for AI

Overview  New
Prepare for setup
Add AI providers
Add integrations
Review data and insights in Fusion EPM
Explore the product
Contact us to get started

API Keys

OCI API keys
Create an API key
Add key permissions

Agent tools

MCP tools
Code interpreter
File search
Enterprise AI NL2SQL
Function Calling

Vector Stores for Enterprise AI features

Create a vector store
Use the NL2SQL API for structured data
Sync data for unstructured data

Community

AI at Oracle
Oracle AI blog
AI solutions hub