Cloudflare has started with an honest description of how agents read the web today. In Introducing Web Search API via AI Gateway, the post explains that when an agent needs a live page, it “usually just guesses the URL of the page and then makes a tool call to curl it,” which is why fetches come back 404 when the guess is wrong. The fix is the one humans use: search first. Cloudflare is launching web search through AI Gateway with three partners, Ceramic.ai, Exa and Linkup, returning “fresh, structured snippets from the web injected straight into context.”
There are no usage figures, and the condition attached to the partners is the part worth keeping. Each partner’s crawler “must comply with Cloudflare’s publicly stated requirements for ‘Verified bots’,” and every web search response “must include a link to the location of crawled content.” That is a supply-side rule enforced at the point of sale: an agent buying search through Cloudflare buys it only from crawlers that identify themselves and respect site preferences. The commercial design follows AI Gateway’s: search draws down gateway credits at partners’ list prices “without any additional markup,” shows up in the same observability logs, supports bring-your-own-key, and flags partners with zero data retention. It is reachable as a REST endpoint under the account’s AI path or as a one-line Workers binding, and Cloudflare says web search will be one of the first built-in server tools in the gateway.
The catalog holds the control plane this lands in. The Cloudflare provider page lists 650 API pages, and the release extends the Cloudflare AI Gateway API, with the Workers binding reached through the Cloudflare Workers API. The agentic access profile maps 3,118 operations, 1,576 of them acting. Sixteen Agent Readiness dimensions are lit.
The Kin Score is 81.3, exemplar band, and Agent Readiness is 74.8, agent-native. This release touches one of the three dimensions still unlit on the record. Agentic commerce measures whether an agent can pay for a capability through the API itself, and here an agent’s search spend is drawn from a gateway balance at a published per-call price, attributed and logged. That is close to the shape the dimension describes, metered and agent-initiated, though still bound to an account the human funds. Cloudflare has made the honest crawler the only crawler an agent can buy from. It has also built most of a meter for what agents spend.