Amazon Bedrock gets built-in Web Search for grounded model responses

AWS announced general availability of Web Search on Amazon Bedrock, a built-in server-side tool that lets foundation models — including OpenAI's GPT-5.4, GPT-5.5, and GPT-5.6 Sol/Terra/Luna — ground their responses in live web knowledge without integrating a third-party search provider. The capability is activated with a single API parameter, returns structured citations, and runs entirely inside the customer's secured AWS environment with zero data egress, preserving data residency.

AWS pitched the feature as removing the operational burden of onboarding external vendors, orchestrating separate APIs, or running additional third-party security reviews just to add web grounding. It expands on the web-search capability first introduced within Bedrock AgentCore in June 2026, promoting it to a first-class native tool across the platform.

The launch fits AWS's broader push to make Bedrock the enterprise control plane for agentic AI, and it landed the same day as a Superblocks–AWS collaboration bringing governed enterprise AI app development into customers' AWS environments via Bedrock and Aurora. For enterprises wary of sending prompts and retrieved content to outside search APIs, the zero-egress guarantee is the differentiating selling point. The competitive context: Google and Microsoft already bundle grounded search into their model stacks, so AWS is closing a parity gap while leaning on its data-governance story. Watch for pricing details and how citation quality compares to dedicated search-grounding vendors.