Skip to content

AWS Lets Agent Builders Restrict Web Search to Approved, Recent Sources

Short answer

AWS added domain and publish-date filters to the Web Search tool in Bedrock AgentCore. Agent builders can now restrict searches to approved domains and require recent content, reducing the chance an AI agent cites outdated, unreliable, or off-brand sources when answering customer or internal queries.

What this means for operators

For a 10-200 person B2B company running a support or sales agent that pulls live web results to answer customer questions, this closes a real gap: until now, an agent grounded in open web search could just as easily surface a three-year-old blog post or a competitor's page as your own documentation. Teams building on AgentCore can now lock search to a whitelist (docs.yourcompany.com, trusted partner sites, industry standards bodies) and require content published within a set window, which matters for anything involving pricing, compliance, or product specs that change often. It also gives ops and legal teams a concrete control to point to when a customer or auditor asks how the agent decided what to cite, rather than an unverifiable 'it searched the web.'

AWS has added domain and publish-date filtering to the Web Search tool inside Amazon Bedrock AgentCore, according to the AWS Machine Learning Blog. AgentCore is AWS's managed runtime for building and running AI agents, and its Web Search tool lets those agents fetch live information from the internet rather than relying solely on a model's training data or a fixed knowledge base.

The new filters let developers constrain a search to a specified list of domains and set a recency window, so results outside the allow-list or older than the cutoff are excluded before they ever reach the agent's reasoning step. AWS frames this as a way to improve both relevance and trust: an agent answering a support question can be restricted to a company's own help center plus a short list of vetted vendor sites, and a research agent can be told to ignore anything published before a given date.

This is a narrow but practical change. Open-ended web search grounding has been one of the harder problems for teams deploying customer-facing or internal AI agents, because the model has no inherent way to judge source quality or currency — it will happily cite a stale forum post with the same confidence as an authoritative source. Domain and date filters push that judgment upstream, into a configuration a builder controls directly, rather than leaving it to prompt instructions the model may or may not follow.

For operations teams, the change matters most in three places: customer support agents that need to cite only company-approved documentation, sales or research agents that need current pricing or competitive information rather than dated pages, and any workflow with a compliance requirement to show which sources an AI system was allowed to draw from. Restricting the source pool also narrows the attack surface for prompt injection via malicious web content, since fewer, known domains are reachable.

The feature is specific to AgentCore's Web Search tool and does not change how other retrieval methods, such as vector databases or internal knowledge bases, behave. Teams already running agents on AgentCore with web search enabled can apply the filters through the tool's configuration; AWS's post does not specify pricing changes tied to the update, so cost impact should be treated as unconfirmed pending direct testing.

Source: AWS Machine Learning Blog

Next step

Visibility Analyzer

This is what the Visibility Analyzer measures on a real site: which answers cite you, which pages an engine cannot retrieve, and what to fix first. Free to run.

Run a free visibility audit

Free to run. No card.

Fee
Free
Length
One run, minutes

Free tier: two analyses a day, no card required.