Build memory-aware AI agents with Agentic Retrieve on Amazon Bedrock Managed Knowledge Base
Amazon Bedrock Managed Knowledge Base and Amazon Bedrock AgentCore memory now store conversations, extract facts, and bring long-term memory into a single AgenticRetrieveStream call. This post shows how to build a memory-aware, multi-user assistant where each user sees only the documents they may read and each user's memory stays private.
First seen by Oriel on .
AI summary
This AWS walkthrough shows how Amazon Bedrock Managed Knowledge Base and Amazon Bedrock AgentCore memory combine to deliver cited, ACL-filtered answers that retain private per-user context.
- Adding memoryConfiguration to AgenticRetrieveStream lets the service restore short-term conversation state, retrieve relevant long-term facts, generate grounded answers, and save each turn.
- The application derives userContext.userId from verified Cognito email for document ACLs, while Cognito sub becomes actorId for stable memory isolation.
- A single shared memory resource uses /anycompany/{actorId} namespaces, with summary and user-preference strategies covering conversational facts and response preferences.
- The managed knowledge base filters documents by ACLs and streams citations and trace events; the sample indexed 11 documents with zero failures.
- The integration is opt-in and backward compatible, requires the knowledge base and memory in the same account and us-west-2, and needs Python 3.10 or later with Boto3 1.43.72 or later.
#amazon bedrock #agentcore memory #knowledge bases #amazon cognito
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| AWS Machine Learning Blog | 6d1ff2e4339ff5c12febb03d6586784822aa87ad | https://aws.amazon.com/blogs/machine-learning/build-memory-aware-ai-agents-with-agentic-retrieve-on-amazon-bedrock-managed-knowledge-base/ Read discussion |