AWS S3 Tables MCP

Local · stdio

AWS S3 Tables is a specialized analytical table format (Apache Iceberg) built on S3, and the official MCP server from the awslabs/mcp monorepo gives agents table-bucket and namespace management, table creation and renaming, SQL queries, and creating tables from CSV files. This isn't general access to regular S3 objects, just its analytical table buckets. Auth is via AWS IAM credentials.

Why this server

create_table_from_csv turns a CSV file into an analytical Iceberg table in one call
query_table runs read-only SQL directly against the tabular data in S3 Tables
list_table_buckets and list_namespaces orient the agent within the data storage structure
get_table_metadata shows a table's schema and metadata with no need for the AWS console

Usage examples

Prompt: Create a table from this sales-log CSV and show its metadata

Result: create_table_from_csv turned the file into an Iceberg table in the right namespace, and get_table_metadata returned its schema and settings.

Prompt: Calculate monthly sales totals in the sales table for this year

Result: query_table ran a read-only monthly SQL aggregation directly against the S3 Tables data, with no need to export it to a separate database.

Local server

This MCP runs locally on the user's machine (stdio), so availability isn't tracked — rely on reviews and the rating instead.

Reviews

No reviews yet.