Summary
Cloudflare changed new AI Search instances so they include built-in storage and vector indexes by default, added namespace-level Workers bindings, and later added Wrangler commands for namespace management. Together, the updates make AI Search easier to provision, automate, and operate from code, scripts, and agent workflows.
What changed
Cloudflare AI Search gained built-in storage, namespace-level Workers bindings, cross-instance search APIs, and Wrangler CLI commands for listing, creating, inspecting, updating, and deleting namespaces.
Why it matters
This makes AI Search more programmable in production. Developers can manage retrieval infrastructure at runtime or from a terminal instead of treating search namespaces as static setup artifacts, which is useful for multi-tenant apps, RAG systems, and agent workflows that need to create or update search surfaces dynamically.
Evidence excerpt
Cloudflare says new AI Search instances include built-in storage and vector indexes, namespace bindings can manage instances at runtime, and Wrangler now supports namespace list, create, get, update, and delete commands.