Cf: The Agentic CLI for the Cloudflare API
Cloudflare introduced cf, an agent-first CLI generated from OpenAPI schemas that exposes over 3,000 operations, JSON-default output, and a TypeScript configuration format.
summary is enough Useful announcement for developers using AI agents with Cloudflare, though the core concept is straightforward wrapper tooling around their OpenAPI specs.
summary
Cloudflare is replacing Wrangler with cf, a new command-line tool built from the ground up for AI agents using its Forge API generation pipeline. It expands coverage from Wrangler's roughly 280 commands to over 3,000 operations across the entire Cloudflare API surface, featuring default JSON outputs, a natural language command search, and a TypeScript-based configuration format integrated with Vite. The open beta is available now via npm, with Wrangler slated for deprecation after a transition period.
key points
- Generated via OpenAPI schemas using Cloudflare's Forge pipeline, expanding CLI coverage from 280 to over 3,000 operations.
- Outputs condensed JSON by default to save context tokens for AI agents and includes a natural language command search feature.
- Adopts a type-safe TypeScript configuration format and switches to Vite for local development.
who should read it
Cloudflare developers and engineers building applications with AI coding agents.
discussion positive
consensus
Commenters generally praise Cloudflare's aggressive push into agent-first tooling, though some debate whether TypeScript is an appropriate language for a CLI tool instead of a compiled binary.
pushback
Some engineers point out that agents already work well with Cloudflare via direct REST API calls, questioning the necessity of a dedicated CLI wrapper.
Kudos to cloudflare on this, it looks like a great CLI, I especially like the `cf cli search`.
Related, Matt Pocock's skill for having agents generate an interactive bash script for things only humans can do (or should do), presented in the context of devops like activities
https://github.com/mattpocock/skills/blob/main/skills/produc...
I have found that having the agents write scripts to use tools like this is better (less tokens, more reliability, fewer side quests) than giving it to them directly with markdown they may or may not follow on any given day.
The other benefit to this is …
Notes that while CLIs like this are useful for read-only debugging, production infrastructure modifications should still rely on proper Infrastructure as Code rather than direct agent execution.
Generated by an LLM from the article and thread. Check the sources before relying on it.