This article explains how to use Kiro, an AI agent, with Amazon Redshift's Model Context Protocol (MCP) server to query data warehouses using natural language instead of SQL.

• Amazon Redshift MCP server bridges Kiro AI with Redshift infrastructure for automated cluster discovery
• Enables metadata exploration, safe read-only query execution, and multi-cluster support
• Installation requires Python 3.10+, uv package manager, and proper IAM permissions configured
• Kiro discovers clusters, databases, schemas, and tables without manual endpoint configuration
• Supports cross-cluster comparisons, data quality checks, and schema documentation generation
• Read-only enforcement prevents accidental writes; least-privilege IAM policies provide primary defense
• Kiro shows generated SQL queries for verification and auditability
• Reduces time spent on infrastructure navigation and SQL syntax debugging

Kiro with Redshift MCP eliminates friction between asking data questions and getting answers, enabling business users to perform complex analysis without SQL expertise or data engineering support.