AIOZ Storage Vector Database: Store, Embed, and Retrieve by Meaning

AIOZ Storage Vector Database is now LIVE.

Create a vector bucket, store embeddings, and retrieve relevant information through semantic similarity, all within AIOZ Storage.

For text workflows, built-in server-side embedding adds another step directly into the experience: send text, let AIOZ infrastructure generate the embedding, and search by meaning.

Semantic Retrieval in AIOZ Storage

Vector databases give AI applications a way to retrieve information based on meaning.

An embedding model converts content into numerical representations called vectors. Semantically similar information is positioned closer together, allowing queries to retrieve relevant results even when the wording is different.

AIOZ Storage brings this model into S3-compatible vector buckets, connecting vector storage and retrieval with familiar storage workflows.

Key capabilities include:

• Meaning-Based Retrieval: Find related information through nearest-neighbor similarity search.
• Built-In Text Embedding: Send text directly and generate embeddings on AIOZ infrastructure.
• Familiar Authentication: Secure requests with AWS Signature Version 4 signing.
• Standard SDK Compatibility: Manage vector buckets through familiar AWS SDKs and tooling.
• Integrated Storage Workflow: Add vector retrieval alongside the broader AIOZ Storage environment.

From Text to Semantic Retrieval

The workflow is straightforward:

1. Create a vector bucket in AIOZ Storage.
2. Add the text you want to store and retrieve.
3. AIOZ infrastructure generates the embedding server-side.
4. Search using a text query.
5. Retrieve the nearest results based on semantic similarity.

What You Can Build

Vector retrieval can support a wide range of AI and application workflows, including:

• Retrieval-Augmented Generation: Retrieve relevant documents to provide context for LLM responses.
• Semantic Search: Surface results based on meaning rather than exact wording.
• Recommendations: Identify items that are semantically similar to one another.
• Agent Retrieval: Recall relevant information based on context and meaning.
• Knowledge Search: Retrieve information across documentation, support content, product data, and other text collections.

Built Around Familiar Storage Tools

A vector bucket stores embeddings within the broader AIOZ Storage environment.

Requests use AWS Signature Version 4, while standard AWS SDK compatibility keeps vector bucket management aligned with familiar developer tooling.

Pay-As-You-Go Vector Storage

AIOZ Storage Vector Database pricing includes:

• Vector storage: $0.05 per GB-month
• Vector upload with embedding included: $0.04 per 1 million tokens
• Query / retrieval: $2.00 per 1 million queries
• Data processed: $0.0028 per TB below 100,000 vectors in an index, scaling to $0.0002 per TB above 10 million vectors
• Data returned: $10.00 per TB

The model separates storage, embedding, retrieval, processing, and returned-data usage, giving teams visibility into each part of the vector workflow.

The Foundation for AIOZ Agent Memory

AIOZ Storage Vector Database also provides the storage layer for the upcoming AIOZ Agent Memory feature.

Agent Memory builds a purpose-designed agent workflow on top of vector storage: record information, then recall it later by meaning.

It uses a fixed embedding model, removes index configuration from the agent-facing workflow, and is designed to return stored content verbatim.

Vector Database remains the general-purpose vector storage and retrieval layer underneath that experience.

Start with a vector bucket

Create a vector bucket, add text, and run your first semantic search.

No separate text-embedding call. No separate vector authentication setup.

Just a familiar storage workflow extended for meaning-based retrieval.

Build your next retrieval workflow with AIOZ Storage Vector Database.