In collaboration with

Retrieval Augmented Generation (RAG) grounds your model in your data. Agentic memory carries context across sessions. Semantic search matches intent, not keywords.

Each puts different demands on the same layer: a vector database. What to evaluate:

• Pattern-specific diagrams: Each architecture with Amazon Bedrock and Amazon Bedrock AgentCore
• Comparison framework: Hybrid search, scaling, hosting models, and metadata filtering
• Free trials in AWS Marketplace: Test popular vector databases against your workload

Choose your vector database by architecture, not by feature list.

Get the evaluation guide ›

About AWS Marketplace:

Visit AWS Marketplace for help finding the best solutions for your use case and add new capabilities to your tech stack with your AWS account:

• Tools are designed to plug in to your AWS workflows and integrate with your favorite AWS services
• Evaluate fast with free trials and developer-tier pricing to support fast prototyping
• Avoid upfront license fees and pay only for what you use, consolidating billing with your AWS account
• A broad selection of tools across AI, data, observability, security and more can enhance how you build with AWS