MySQL AI User Guide

Abstract

This document describes how to use MySQL AI. It covers how to load data, run queries, optimize analytics workloads, and use machine learning and generative AI capabilities.

For legal information, see the Preface and Legal Notices.

For help with using MySQL, please visit the MySQL Forums, where you can discuss your issues with other MySQL users.

Document generated on: 2025-09-10 (revision: 83501)

Table of Contents

Preface and Legal Notices

1 Introduction to MySQL AI

2 Installing MySQL AI

2.1 Supported Platforms and Requirements

2.2 Installing MySQL AI

2.2.1 MySQL AI GUI Installation

2.2.2 Command-line Installation

3 Loading Data in MySQL AI

3.1 Bulk Ingest Data

3.2 Bulk Ingest Data to MySQL Server Limitations

4 Training and Using Machine Learning Models

4.1 About AutoML

4.1.1 AutoML Ease of Use

4.1.2 AutoML Workflow

4.1.3 AutoML Learning Types

4.1.4 Oracle AutoML

4.2 Additional AutoML Requirements

4.3 AutoML Privileges

4.4 Supported Data Types for AutoML

4.5 Creating a Machine Learning Model

4.5.1 Preparing Data

4.5.2 Training a Model

4.5.3 Loading a Model

4.5.4 Generating Predictions

4.5.5 Generating Model Explanations

4.5.6 Generating Prediction Explanations

4.5.7 Scoring a Model

4.6 Machine Learning Use Cases

4.6.1 Classify Data

4.6.2 Perform Regression Analysis

4.6.3 Generating Forecasts

4.6.4 Detect Anomalies

4.6.5 Generating Recommendations

4.6.6 Generating Topic Modeling

4.7 Manage Machine Learning Models

4.7.1 The Model Catalog

4.7.2 Work with Model Handles

4.7.3 Unload a Model

4.7.4 View Model Details

4.7.5 Delete a Model

4.7.6 Share a Model

4.7.7 Manage External ONNX Models

4.7.8 Analyzing Data Drift

4.8 Monitoring the Status of AutoML

4.9 AutoML Limitations

5 AI-Powered Search and Content Generation

5.1 About GenAI

5.2 Additional GenAI Requirements

5.3 Required Privileges for using GenAI

5.4 Supported LLM, Embedding Model, and Languages

5.5 Generating Text-Based Content

5.5.1 Generating New Content

5.5.2 Summarizing Content

5.6 Setting Up a Vector Store

5.6.1 About Vector Store and Vector Processing

5.6.2 Ingesting Files into a Vector Store

5.6.3 Updating a Vector Store

5.7 Generating Vector Embeddings

5.8 Performing Vector Search with Retrieval-Augmented Generation

5.8.1 Running Retrieval-Augmented Generation

5.8.2 Using Your Own Embeddings with Retrieval-Augmented Generation

5.9 Starting a Conversational Chat

5.9.1 Running GenAI Chat

5.9.2 Viewing Chat Session Details

6 Review Performance and Usage

6.1 MySQL AI Performance Schema Tables

6.1.1 The rpd_column_id Table

6.1.2 The rpd_columns Table

6.1.3 The rpd_ml_stats Table

6.1.4 The rpd_nodes Table

6.1.5 The rpd_preload_stats Table

6.1.6 The rpd_table_id Table

6.1.7 The rpd_tables Table

6.2 Option Tracker

7 MySQL AI Routines

7.1 AutoML Routines

7.1.1 ML_TRAIN

7.1.2 ML_EXPLAIN

7.1.3 ML_MODEL_EXPORT

7.1.4 ML_MODEL_IMPORT

7.1.5 ML_PREDICT_ROW

7.1.6 ML_PREDICT_TABLE

7.1.7 ML_EXPLAIN_ROW

7.1.8 ML_EXPLAIN_TABLE

7.1.9 ML_SCORE

7.1.10 ML_MODEL_LOAD

7.1.11 ML_MODEL_UNLOAD

7.1.12 ML_MODEL_ACTIVE

7.1.13 Model Types

7.1.14 Optimization and Scoring Metrics

7.2 GenAI Routines

7.2.1 ML_GENERATE

7.2.2 ML_GENERATE_TABLE

7.2.3 VECTOR_STORE_LOAD

7.2.4 ML_RAG

7.2.5 ML_RAG_TABLE

7.2.6 HEATWAVE_CHAT

7.2.7 ML_EMBED_ROW

7.2.8 ML_EMBED_TABLE

8 Troubleshoot

8.1 AutoML Error Messages

8.2 GenAI Issues