Artificial intelligence news.
Straight from the original announcements.

AWS adds coding-agent guide for Amazon Bedrock knowledge base migrations

November 12, 2026, 17:16By the AI News Bites Editorial TeamSource: AWS
AWS adds coding-agent guide for Amazon Bedrock knowledge base migrations
Image: AWS

AWS has added a migration guide to the Agent Toolkit for AWS to help teams move production customer-managed knowledge bases to Amazon Bedrock Managed Knowledge Base. The guide uses a coding agent to map configurations, plan changes and create the replacement alongside the original.

Key points
  • AWS added a knowledge base migration guide to the Agent Toolkit for AWS.
  • The coding agent starts with read-only discovery and a reviewable migration plan.
  • Managed Knowledge Base handles storage, indexing, ingestion and retrieval.
  • Managed knowledge bases support 200 data sources by default.
  • Managed Knowledge Base charges $1 per 1,000 standard retrieval requests.

Managed Knowledge Base takes over infrastructure work

With a customer-managed knowledge base, teams provision and scale the vector store, tune the embedding model and index settings, and plan capacity for each data source. Managed Knowledge Base moves ingestion, storage, indexing and retrieval into Amazon Bedrock, with automatic scaling as content grows.

AWS says the service combines keyword and semantic matching through always-on hybrid search, without requiring teams to configure an index. It also includes a managed embedding model, a managed reranker and a built-in parser for multimodal file types. Teams can still select their own Amazon Bedrock embedding model when creating the knowledge base.

Managed Knowledge Base also offers native connectors for sources including Amazon S3, Microsoft SharePoint and Atlassian Confluence. Later synchronizations can process only changed content, either on demand or on a schedule. Other capabilities include a deletion safeguard, access control list awareness, agentic retrieval, AgentCore Gateway integration and resource-based policies.

The migration keeps the original knowledge base running

AWS describes the migration as a parallel process. The coding agent creates a new MANAGED knowledge base alongside the original VECTOR knowledge base, without modifying or deleting the original during migration.

The guide begins with read-only discovery and produces a plan for the team to review line by line. The plan identifies settings that remain unchanged, differences handled by Managed Knowledge Base and application changes required for the migration.

After approval, the agent can create the new knowledge base and its data sources alongside the original, or write a migrated infrastructure as code template for the team’s pipeline. The team remains responsible for ingesting content, updating the application’s Retrieve calls and validating retrieval before switching traffic.

Higher quotas and usage-based pricing shape the decision

AWS lists higher default quotas for Managed Knowledge Base. A managed knowledge base supports 200 data sources, compared with 5 for a customer-managed knowledge base, while the account and Region limit is 10,000 managed knowledge bases compared with 100 customer-managed ones. The concurrent ingestion job limit is 50 instead of 1, and query input characters rise from 1,000 to 10,000.

The pricing model also differs. AWS says Managed Knowledge Base charges for what is stored and retrieved, with vector storage included in a base rate of $5 per GB per month and standard retrieval priced at $1 per 1,000 retrieval requests. Agentic retrieval costs an additional $4 per 1,000 queries on top of underlying retrieval calls.

For users, the guide is designed to reduce migration planning and operational work while preserving control over content and the application. AWS presents the move as a trade-off between ease of use, retrieval capabilities, cost and the direct control available with a customer-managed vector store.

AWSAmazon Bedrock Managed Knowledge BaseAgent Toolkit for AWSAmazon Simple Storage ServiceMicrosoft SharePointAtlassian Confluenceenterprise AIcloudAI agents

Related news

AWS details four ways to speed up Amazon Bedrock AgentCore sessions
For developersNovember 13, 2026, 23:02

AWS details four ways to speed up Amazon Bedrock AgentCore sessions

AWS has outlined four architecture patterns for reducing session startup delays in Amazon Bedrock AgentCore runtime, including snapshot-based initialization, build hooks…

AWS upgrades AgentCore runtime with faster starts and stronger isolation
Products and toolsNovember 12, 2026, 18:59

AWS upgrades AgentCore runtime with faster starts and stronger isolation

AWS has announced updates to the Amazon Bedrock AgentCore runtime, giving AI agent sessions faster startup, more compute capacity and expanded isolation for data. The…

AWS launches direct Web Search API for Bedrock AgentCore
Products and toolsNovember 12, 2026, 00:13

AWS launches direct Web Search API for Bedrock AgentCore

AWS has introduced the Web Search API on Amazon Bedrock AgentCore, allowing applications to retrieve ranked web results through a single API call. The service is…