Google Cloud adds object enrichments for agent-ready data

Google Cloud has introduced object enrichments, a Cloud Storage capability that uses AI models to extract contextual metadata from unstructured files and store it alongside the source objects. The preview service is designed to help AI agents work with images, PDFs, videos and audio without repeatedly processing the raw files.
- Object enrichments is available in preview for Google Cloud Storage.
- It stores AI-generated metadata as object context alongside source objects.
- Google Cloud says enriching 10 million images or PDFs costs $700.
- The service includes five first-party models for images and PDFs.
- Gemini support is planned for advanced multimodal processing.
Turning stored files into usable context
Google Cloud says most enterprise information is unstructured and stored as petabytes of images, PDFs, videos and audio files. While this data can contain knowledge needed by AI agents, a raw object appears to an agent as bytes that it cannot directly reason over.
Object enrichments is intended to address that problem by automatically extracting contextual metadata and saving it as object context in Cloud Storage. Google Cloud says the capability can provide image annotations such as object detection, face detection, image labeling and explicit content detection, as well as PDF classification across categories including Finance, Games, News and Health.
The company says the service uses models also used by internal Google services such as Google Photos. It claims image annotations are available at 75% cost savings compared with similar services from other cloud providers.
Lower-cost processing without custom pipelines
Google Cloud says enriching 10 million images or PDFs costs $700, according to published pricing. The service is managed through the Google Cloud console without code, or programmatically through the Cloud Storage API.
Customers select buckets and a processing location, choose one or more of five first-party Cloud Storage models for images and PDFs, and grant the required IAM permissions. Object enrichments then handles model inference, response parsing, retries and throttling. Google Cloud says customers do not need to manage queues, scale workers or maintain glue code.
For more complex multimodal tasks involving text, PDFs, images, audio and video, customers can use Gemini with object enrichments for applications such as summarization, entity extraction, transcription and translation. Google Cloud says this upcoming functionality requires customers to request access.
Metadata stays with the source object
Object context is a Cloud Storage metadata construct that stores structured key-value pairs alongside an object's bytes. According to Google Cloud, the context shares the underlying object's durability and consistency properties, moves with the object during copy, move and replication operations, and is removed when the object is deleted.
IAM controls can govern who reads, adds, updates or removes the context. Applications can also filter object listings using context key-value pairs or read the metadata alongside objects. Google Cloud says object context can flow into BigQuery through Storage Insights datasets, allowing teams to run SQL analytics and join insights from unstructured files with structured enterprise tables without writing ETL code.
The preview is aimed at use cases including creative asset search, content moderation and training-data curation. Google Cloud cites L'Oreal, Hyperscience and Vivint as early customers. L'Oreal said the capability helped label its creative asset repository “100x faster at a fraction of the cost of existing pipelines and processes,” while Hyperscience reported a “50%+ lift” in document classification accuracy and a 2-3x reduction in classification inference costs.


