What changed

Amazon Bedrock AgentCore has introduced the Web Search API, which enables agents and applications to retrieve current web results through a single API call.

This API is signed with the AWS credentials that users already have, eliminating the need for a separate search API key. The API is served from a web index operated by Amazon, rather than being sent to a third-party search engine.

Each result includes the page’s title, URL, publication date, and an extract relevant to the query. Users can include or exclude specific domains and limit results to a publication date range.

What this means in practice

The Web Search API is available on AWS Lambda, Amazon Elastic Container Service (Amazon ECS), and Amazon Bedrock AgentCore runtime.

Agents can call the API using their execution role, and the results are returned as data that can be checked by the code.

The API is also accessible through AgentCore Gateway, where one administrator source policy covers all agents and Model Context Protocol (MCP) clients. For LangChain developers, the AgentCore web search tool now calls the Web Search API, providing current, cited search results.

The tool can also enforce rules such as a date window, a source blocklist, and a search budget, which the model cannot change.

Limitations and availability

The Web Search API is available in the US East (N. Virginia) Region (us-east-1) and other supported AWS Regions. It is billed per use, and one run of the watch makes at most 12 searches.

The complete code is in a notebook, which runs every cell, including setup.

The API requires AWS credentials and specific IAM permissions, such as the bedrock-agentcore:InvokeWebSearch action on the resource arn:aws:bedrock-agentcore:us-east-1:aws:tool/web-search.v1. The API does not send queries to third-party search engines and serves results from Amazon’s own web index.

Source facts for Amazon Bedrock AgentCore Adds Web Search API for Real-Time Data: A model’s training data can’t answer those questions, so agents need web search. For production use, teams also need results they can cite, and control over which sources count. They want search queries to stay inside their cloud environment because those queries often reveal internal work.
Original AINewsTrend explainer based on the cited source. This graphic summarizes source statements; it is not an independent benchmark.

Sources and further reading

  1. awsml: Build a daily dependency watcher with LangGraph and Amazon Bedrock AgentCore web search