ReleaseBytes
aws AWS News Blog ·

Amazon S3 adds annotations for rich, queryable object metadata

blogdataawsengineeraws-s3
feature

Amazon S3 now supports annotations, allowing users to attach up to 1 GB of rich, mutable metadata per object in flexible formats like JSON, XML, or plain text. This feature is designed to support AI agents and autonomous workflows by providing scalable, queryable context that moves with the object. When S3 Metadata annotation tables are enabled, annotations are automatically indexed into queryable Apache Iceberg tables accessible via Amazon Athena. This aims to simplify complex metadata management across industries by eliminating the need for separate databases or sidecar files.

  • →New S3 annotations provide large-scale, queryable object context
  • →Query annotations at scale with S3 Metadata annotation tables
  • →Annotations offer greater flexibility than existing metadata options
  • →Annotations support AI agents and autonomous workflows
  • →Manage annotations using S3 API actions
Features (2) ›
  • New S3 annotations provide large-scale, queryable object context

    Amazon S3 annotations allow attaching up to 1,000 named annotations per object, totaling 1 GB, in formats like JSON, XML, YAML, or plain text. This metadata is mutable and moves with the object during transfers, simplifying context management for AI agents and workflows. Annotations can be queried at scale when S3 Metadata annotation tables are enabled, automatically indexing into Apache Iceberg tables accessible via Amazon Athena.

  • Query annotations at scale with S3 Metadata annotation tables

    Enabling S3 Metadata annotation tables automatically indexes object annotations into a fully managed Apache Iceberg table. This allows querying across all annotations using Amazon Athena or other Iceberg-compatible engines. Annotations are queryable for objects in any storage class without requiring object restoration or retrieval charges. The annotation tables refresh within approximately one hour.

Enhancements (1) ›
  • Annotations offer greater flexibility than existing metadata options

    S3 annotations provide a more scalable and flexible approach to metadata compared to system-defined metadata, object tags, and user-defined metadata. Annotations support larger data sizes (up to 1 GB per object) and are mutable, unlike user-defined metadata which is set at upload. This addresses limitations for use cases requiring rich, evolving context without managing separate metadata systems.

Maintenance (1) ›
  • Manage annotations using S3 API actions

    Users can manage S3 annotations using API actions such as PutObjectAnnotation to add or update metadata, GetObjectAnnotation to retrieve specific annotations, ListObjectAnnotations to see all attached annotations, and DeleteObjectAnnotation to remove them. IAM policy or bucket policy permissions for s3:PutObjectAnnotation and s3:GetObjectAnnotation are required.

Notes (1) ›
  • Annotations support AI agents and autonomous workflows

    The new annotation feature enables AI agents and autonomous workflows to find and act on data by providing rich, queryable context directly with objects. This supports use cases like AI-generated transcripts, content ratings, or technical specifications. Annotations automatically flow with objects during copy, replication, and cross-region transfers, and are removed upon object deletion.

Read the original announcement →

https://aws.amazon.com/blogs/aws/amazon-s3-annotations-attach-rich-queryable-context-directly-to-your-objects/

Related releases