Cloudflare has brought native web search into AI Gateway, enabling developers to ground model responses in real-time web context without wiring up a separate retrieval pipeline. The capability arrives through partnerships with Ceramic.ai, Exa, and Linkup, giving teams a choice of search providers behind a single Cloudflare interface.
Key changes announced in the release:
- AI Gateway now supports native web search API integration, exposing search as a first-class capability rather than a bolt-on tool.
- Three search partners are available at launch: Ceramic.ai, Exa, and Linkup.
- Developers can inject real-time web context into model inference calls through AI Gateway, the REST APIs, or Workers bindings.
- The integration is reached via the existing
AI Gatewaysurfaces, so search results can be passed alongside prompts in the same request flow.
The design targets the common grounding problem: language models are trained on fixed data and can go stale between releases. By routing search through AI Gateway, developers avoid standing up their own crawling, indexing, and result-merging infrastructure, and they keep provider credentials and configuration in one place. The Workers bindings option is particularly relevant for edge applications, where a Worker can fetch fresh results and pass them into an inference call without an extra round trip to an external orchestration service.
For developers, the practical impact is faster time-to-production for retrieval-augmented features such as current-events answers, citation-backed summaries, and agentic workflows that need up-to-date facts. Because the API is exposed both as a REST endpoint and as a binding, the same integration can serve backend services and edge-deployed Workers. Teams already using AI Gateway can add web context without changing their underlying model provider, while those comparing search vendors can swap between Ceramic.ai, Exa, and Linkup behind a consistent interface.
Source: https://blog.cloudflare.com/introducing-web-search-api/