Does Cloudflare's server manage a full cloud like AWS?

No. Its tools cover the edge and serverless layer: accounts (accounts_list, set_active_account), Workers KV, Workers inspection, and R2 buckets. For broad compute, IAM, and networking, AWS (AWS Labs) is our top pick, with Google Cloud Run and Azure as the other clouds.

Which Cloudflare resources can the agent actually touch?

Through the listed tools: KV namespaces (list, create, get, update, delete), Workers (list, get details, get code), and R2 buckets (list, create), scoped by accounts_list and set_active_account. It is best paired with a major-cloud server for the rest of the infrastructure.

Cloudflare for cloud infrastructure

Pick 4 of 4 for cloud infrastructureOfficialCloudflare3,806

Cloudflare's remote MCP servers let an agent build and manage Workers, KV, R2, D1, and Hyperdrive, the edge and serverless layer many teams put in front of their cloud. For cloud infrastructure broadly it is our fourth pick of four, because the job here is operating the full compute, storage, networking, and database footprint of a cloud, and Cloudflare covers the edge slice rather than that whole surface.

The honest read: Cloudflare is the right server when the infrastructure question is about the edge, but a partial answer for the cloud as a whole. It ranks fourth on a list led by the major clouds, while still being the natural choice for the serverless layer that fronts them.

How Cloudflare fits

The exposed tools center on accounts, edge key-value, compute, and object storage. accounts_list and set_active_account scope which account tool calls run against. kv_namespaces_list, kv_namespace_create, kv_namespace_get, kv_namespace_update, and kv_namespace_delete manage Workers KV. workers_list, workers_get_worker, and workers_get_worker_code inspect deployed Workers and their source. r2_buckets_list and r2_bucket_create handle R2 object storage. An agent can survey and stand up the edge resources directly.

The limits matter against the siblings. There are no tools here for general VMs, IAM, networking, or the broad service catalog a full cloud needs, and the listed D1 and Hyperdrive operations are not present as tools even though the platform covers them. AWS (AWS Labs), our top pick, spans that wide infrastructure surface; Google Cloud Run covers serverless containers on GCP; Azure covers the Microsoft cloud. Use Cloudflare when the agent's job is the edge and serverless layer, and pair it with one of those for the rest of the stack.

Tools you would use

Tool	What it does
accounts_list	List all accounts in your Cloudflare account.
set_active_account	Set the active account used for tool calls that require an account ID.
kv_namespaces_list	List the Workers KV namespaces in your account.
kv_namespace_create	Create a new Workers KV namespace.
kv_namespace_delete	Delete a Workers KV namespace.
kv_namespace_get	Get the details of a Workers KV namespace.
kv_namespace_update	Update the title of a Workers KV namespace.
workers_list	List the Workers deployed in your account.
workers_get_worker	Get the details of a Cloudflare Worker.
workers_get_worker_code	Get the source code of a Cloudflare Worker (may be the bundled version).

Full Cloudflare setup and config →

FAQ

Does Cloudflare's server manage a full cloud like AWS?: No. Its tools cover the edge and serverless layer: accounts (accounts_list, set_active_account), Workers KV, Workers inspection, and R2 buckets. For broad compute, IAM, and networking, AWS (AWS Labs) is our top pick, with Google Cloud Run and Azure as the other clouds.
Which Cloudflare resources can the agent actually touch?: Through the listed tools: KV namespaces (list, create, get, update, delete), Workers (list, get details, get code), and R2 buckets (list, create), scoped by accounts_list and set_active_account. It is best paired with a major-cloud server for the rest of the infrastructure.