Summary
Cloudflare added a new AI Gateway REST API on api.cloudflare.com that lets developers call OpenAI, Anthropic, Google, and Workers AI models through one authentication model and one set of compatible endpoints. The release turns Cloudflare's earlier AI-platform packaging into a more concrete drop-in control plane for multi-model agent workloads, with logging, caching, rate limiting, guardrails, and unified billing applied automatically.
What changed
Cloudflare expanded AI Gateway with a unified REST API surface that supports provider-agnostic model calls through OpenAI- and Anthropic-compatible endpoints on api.cloudflare.com.
Why it matters
This is a substantive follow-through on Cloudflare's unified inference thesis. Instead of only promising one control plane across providers, Cloudflare now offers a practical API layer that can sit in front of third-party and first-party models while preserving gateway controls such as caching, guardrails, and billing visibility.
Evidence excerpt
Cloudflare says AI Gateway now uses the AI REST API on api.cloudflare.com, with unified endpoints for chat/completions, responses, and messages, while logging, caching, rate limiting, guardrails, and unified billing apply automatically.