AWS Launches Web Search for Bedrock AgentCore, Letting AI Agents Stay Current Without Leaking Data

Amazon Web Services has announced the general availability of Web Search on Amazon Bedrock AgentCore, a fully managed tool that lets AI agents pull in current information from the web while keeping data inside the customer's secured AWS environment.

According to the AWS Machine Learning blog post, dated June 19, 2026, the feature allows AI agents to ground their responses in up-to-date, cited web knowledge — with zero data egress from the customer's AWS setup. The companion AWS News blog announcement appeared a couple of days earlier, around June 17, 2026.

The headline benefit is simple: AI agents that would otherwise rely only on the information they were trained on can now reference what's happening right now, and point to where the information came from.

Web Search plugs into AgentCore through a built-in connector target on the Bedrock AgentCore Gateway, using the Model Context Protocol (MCP).

In practice, an agent sends a natural-language query, and Web Search returns the most relevant snippets, source URLs, titles and publication dates. The model can then reason over those results to produce a grounded, traceable response.

What stands out for the people building these systems is how little plumbing is involved. As AWS puts it, there are no search APIs to provision, no outbound credentials to manage, and no result-parsing glue to maintain. That removes a chunk of the integration work that has typically come with bolting web access onto an AI agent.

A key distinction AWS draws is that many existing "add web search to your agent" solutions are simply wrappers around a third-party search engine. Web Search on Bedrock AgentCore is instead backed by a web index that Amazon operates directly, spanning tens of billions of documents.

Amazon says it refreshes the index on an ongoing basis, reflecting new content within minutes. The tool is built on Amazon's existing search infrastructure, informed by years of experience powering agentic search across products including Alexa+ and Kiro.

Grounding goes beyond plain web pages too. The tool uses a multi-source approach that combines the web index with structured knowledge graph data — adding verified facts and real-time information such as stock prices and sports scores. For businesses that need tighter control, AWS also lets you restrict which domains the tool can query by configuring a domain denylist.

At launch, Web Search on Amazon Bedrock AgentCore is generally available in a single region — US East (N. Virginia) — so it is not yet available everywhere.

Pricing is usage-based. Customers are charged on the number of search queries their agents submit, at $7 per 1,000 queries. New AWS customers also receive up to $200 in Free Tier credits.

Gen Digital, the company behind Norton, was cited as an early user. "With the Web Search tool on Amazon Bedrock AgentCore, Norton Revamp helps professionals build their online reputation with current, grounded content ideas shaped by what's actually happening in the world today. What we value most is that AWS uses its own search index and keeps queries within our trusted AWS environment," said Iskander Sanchez-Rola, Senior Director of AI & Innovation, Gen Digital.

It's worth noting that AgentCore itself reached general availability earlier, on October 13, 2025. This Web Search tool is a newer addition to that platform.

The appeal here is two-sided. On one hand, AI agents become far more useful when they can reason over current, cited information rather than a fixed snapshot of the past. On the other, the design keeps queries inside the customer's secured AWS environment — an answer to the data-residency and security concerns that often slow AI adoption in regulated or privacy-sensitive industries.

For an Australian accounting firm, a retailer monitoring competitor pricing, or a healthcare provider checking the latest guidance, the combination of freshness, traceability and data control is exactly the kind of thing that turns AI from an interesting experiment into something you can actually trust in day-to-day work.

Tools like Web Search make AI agents smarter about the outside world — but their value still depends on how well they understand your world. An agent grounded in fresh web data and verified facts is only as good as the internal business data it can pair that knowledge with.

That's where a unified data foundation matters. When your CRM, accounting, operations and other systems feed into one clean, consistent database, AI agents can connect what's happening in the market to what's happening in your own business — confidently and securely.

The steady stream of releases like this is a reminder that the AI layer is advancing quickly. The businesses best placed to take advantage aren't necessarily the ones chasing every new tool, but the ones with their data already organised, accurate and AI-ready when the right capability arrives.