Amazon Bedrock Managed Knowledge Base accelerates enterprise AI with native connectors and advanced retrieval
Amazon Bedrock's new Fully Managed Knowledge Bases simplifies building enterprise RAG pipelines by providing native data connectors, Smart Parsing for automatic multi-format data preparation, and an Agentic Retriever for complex multi-step queries—all integrated with AgentCore Gateway so developers can focus on business outcomes rather than infrastructure management.
Today, AWS announced Amazon Bedrock Managed Knowledge Base, a new capability enabling developers to build enterprise-grade generative AI applications with proprietary data in minutes. It addresses challenges in connecting to disparate enterprise data sources, optimizing retrieval-augmented generation (RAG) accuracy, and managing infrastructure at scale. The service abstracts complex infrastructure components—storage, retrieval, embeddings, re-ranking, and foundation model selection—into a single managed primitive. By default, it automatically manages embeddings, re-ranker, and foundation models, allowing quick setup without manual selection. Key features include six native data connectors (Amazon S3, SharePoint, Confluence, Web Crawler, Google Drive, OneDrive) that natively pull enterprise data and permissions, Smart Parsing which automatically selects optimal parsing strategies for diverse content types, and Agentic Retriever optimized for complex multi-step, multi-hop queries across multiple knowledge bases. Developers can create a Managed Knowledge Base via the Amazon Bedrock or AgentCore consoles with a few clicks, selecting data sources from supported connectors and automatically generated IAM roles. Smart Parsing uses connector-specific data models, multimodal processing, and optimized chunking to ensure accurate data ingestion and retrieval. Agentic Retriever decomposes complex queries into step-by-step plans performing multi-hop retrieval and reasoning, improving accuracy for agentic AI applications. Integration with Amazon Bedrock AgentCore Gateway exposes knowledge bases as pre-built target types with automatic permission management, observability, and compliance with the Model Context Protocol (MCP), enabling compatibility with frameworks like LangChain, LlamaIndex, CrewAI, LangGraph, and Strands Agents. The service preserves flexibility by allowing developers to select from any foundation, embedding, and re-ranking models available on Bedrock, enabling optimization for accuracy and cost without changing infrastructure. Amazon Bedrock Managed Knowledge Base is available in multiple AWS regions including US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney, Tokyo), Europe (Dublin, Frankfurt, London), and AWS GovCloud (US-West). Pricing is pay-as-you-go based on indexed data size and retrievals, with no upfront commitments, and is included in the AWS Free Tier for new customers. This offering enables developers to focus on building generative AI applications without managing complex retrieval infrastructure.