Amazon Bedrock made Web Search available as a built-in tool for model grounding
AWS launched the general availability of Web Search on Amazon Bedrock – a built-in server-side tool that grounds foundation model responses in up-to-date web content through a single API parameter, with data remaining within the AWS environment.
AWS announced the general availability (GA) of Web Search on Amazon Bedrock – a built-in server-side tool that grounds foundation model responses in up-to-date web knowledge. This follows the earlier announcement of Web Search for AgentCore at AWS New York Summit 2026. According to AWS, the solution eliminates the need to find, integrate, and maintain a third-party search provider, which previously delayed projects and introduced data location risks and operational overhead.
According to AWS, the tool combines a web index operated by Amazon, which includes billions of documents and is continually refreshed, with a built-in knowledge graph that anchors entities and the relationships between them. According to the company, for purely factual queries, this graph is intended to answer with high confidence instead of requiring the model to infer the answer from fragments of page text. The tool also performs semantic snippet extraction – instead of passing the entire page to the model, it provides only relevant passages optimized for its context window, which, according to AWS, saves tokens and speeds up responses.
It is enabled through a single parameter in an existing call to the OpenAI-compatible Responses API, without requiring separate API keys, SDKs, or an orchestration layer – authentication uses existing AWS credentials and a short-lived bearer token (valid for up to 12 hours) generated via SigV4. According to the description, Bedrock handles the entire search lifecycle on the server: it formulates the query, retrieves content from the web index and knowledge graph, inserts the results, including snippets, source URLs, and titles, into the model context, and returns a response with structured citations. By default, the tool offers zero data egress outside the AWS environment. At launch, Web Search is available for OpenAI models running through the new-generation Bedrock inference engine; an optional parameter for live web access is already in place in the API, but currently only serves the pre-indexed web corpus. Details can be found in the source article.
Why it matters
Developers no longer need to integrate and maintain an external search provider – according to AWS, grounding in up-to-date web data is enabled with a single parameter directly in an existing API call. According to the provider, the tool offers companies with data residency and security compliance requirements zero data egress outside the AWS environment by default, which may simplify approval processes compared with connecting to a third party.
Two audiences, two different impacts
What this means
For individuals
Developers building applications on Amazon Bedrock can enable model grounding in web data with a single parameter in an existing OpenAI-compatible API call, without having to build their own tool-handling loop, parse external API responses, or handle request limits and retries.
For a business
According to AWS, companies using Amazon Bedrock can eliminate third-party onboarding for web search and the associated vendor security review because data does not leave the AWS environment by default – which is particularly relevant for regulated industries with data residency requirements.
DevelopmentCheck the original
Event sources
clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.