Menu
Mattering Automations connector

A Cloudflare MCP for workflows that reach past Cloudflare

Connect one scoped API token and DNS records, cache purge, zone settings, Workers KV, Worker scripts, R2 bucket management, ruleset audits, Page Rules, and bulk redirect lists become workflow steps. An AI agent can manage DNS with a backup step built in, a deploy webhook can purge exactly the right cache paths, and the results land next to your tasks and databases instead of in a terminal.

56 ready-made nodes One credential to connect Works with Claude, ChatGPT, and Matter

What the Cloudflare MCP connector does

The Cloudflare connector for Mattering Automations is 56 nodes across the Cloudflare API v4: zones and their settings, DNS records with BIND zone-file export and import, cache purge in every mode, Workers KV from single values to 10,000-pair bulk writes, deployed Worker scripts with their settings, R2 bucket management, read-only ruleset audits, Page Rules, and the account lists that power Bulk Redirects. One scoped API token authenticates all of it.

Every node speaks MCP, the open protocol for connecting AI assistants to tools, so Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. An agent can list a zone's DNS records, export the zone file as a backup, point a record at a new origin, and purge the cache for the paths that changed, without anyone opening the Cloudflare dashboard.

Cloudflare publishes its own official MCP servers, and they are good at what they are for: giving an assistant a direct line to Cloudflare. This connector is for the other half of the job, where Cloudflare is one step among many. On the Mattering canvas the same nodes sit next to every other connector and the platform itself, so a deploy webhook can purge cache and then open a verification task, DNS changes can be gated behind a human approval step, and an audit sweep can write what it finds into a Mattering database. Use whichever fits, or both.

The safety modeling is stated plainly. Of the 56 nodes, 30 read and 26 act, and nine are flagged destructive: zone, DNS record, KV namespace, KV value, KV bulk, Worker script, R2 bucket, Page Rule, and list deletes. Some write surfaces are deliberately not shipped at all: Worker code uploads stay in Wrangler or CI, ruleset writes are excluded because one bad call can replace a zone's entire WAF phase, and R2 object reads and writes belong to R2's separate S3-compatible API with its own credentials. What remains is the surface an automation can hold safely.

One connection, three surfaces

  • Visual automations. Drag Cloudflare nodes onto the canvas and wire them to triggers, conditions, and other apps.
  • AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
  • The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.

Every Cloudflare node, ready to drop on the canvas

56 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.

Read and query nodes 30

Cloudflare: Export DNS Records

Export the zone's complete DNS record set as a BIND zone file (plain text at data.content). The natural backup step before bulk changes, and the input format Cloudflare: Import DNS Records accepts.

Cloudflare: Get Account

Fetch one account: name, type, settings, and created date.

Cloudflare: Get Bulk Operation Status

Poll an asynchronous list operation (from Add List Items) until its status is completed or failed.

Cloudflare: Get Current User

Fetch the user the token belongs to: id, email, name, and two-factor status. Requires a USER-owned token with User Details read permission; account-owned tokens get a permission error here (use Verify API Token instead).

Cloudflare: Get DNS Record

Fetch one DNS record: type, name, content, TTL, proxy status, comment, and tags.

Cloudflare: Get KV Key Metadata

Read the JSON metadata stored with a key, without fetching the value itself.

Cloudflare: Get List

Fetch one list's properties: kind, description, item count, and how many rules reference it.

Cloudflare: Get List Items

List a list's items - for a redirect list, each item's source URL, target URL, and status code. Cursor paginated.

Cloudflare: Get Page Rule

Fetch one Page Rule: its URL match pattern, actions, priority, and status.

Cloudflare: Get R2 Bucket

Fetch one R2 bucket's properties: creation date, location, and storage class.

Cloudflare: Get Worker Script

Download a Worker's deployed source code as text at data.content (module Workers return the bundle content). Capped at 200,000 characters in the node output. Uploading new code stays in Wrangler or CI; this connector reads and manages but does not deploy.

Cloudflare: Get Worker Script Settings

Read a Worker's metadata settings: Logpush flag, tags, tail consumers, and observability configuration.

Cloudflare: Get Zone

Fetch one zone: status, type, paused flag, assigned Cloudflare name servers, and account.

Cloudflare: Get Zone Phase Entrypoint

Fetch the zone's entry point ruleset for a phase, with its rules: what actually runs for that phase on live traffic. Returns 404 when the zone has no ruleset in the phase yet.

Cloudflare: Get Zone Ruleset

Fetch one zone ruleset INCLUDING its rules: each rule's action, filter expression, description, and enabled flag. The audit view of a zone's WAF custom rules.

Cloudflare: Get Zone Setting

Read one zone setting by its ID, for example ssl, security_level, cache_level, development_mode, always_use_https, or browser_cache_ttl.

Cloudflare: List Account Lists

List the account's lists (redirect, ip, hostname, asn kinds) with item counts. Redirect lists are the storage half of Bulk Redirects.

Cloudflare: List Account Rulesets

List account-level rulesets, including the Bulk Redirect phase (http_request_redirect) and Enterprise custom rulesets.

Cloudflare: List Accounts

List the Cloudflare accounts the token can access, with their account IDs. Almost every account-level node (KV, Workers, R2, lists) needs an account ID from here.

Cloudflare: List DNS Records

List a zone's DNS records with filters on type, name, content, comment, and proxy status. Returns record IDs used by the get, update, and delete nodes.

Cloudflare: List KV Keys

List key names in a KV namespace, optionally filtered by prefix, with each key's expiration and metadata. Cursor paginated: pass data.result_info.cursor back to get the next page.

Cloudflare: List KV Namespaces

List the account's Workers KV namespaces with their IDs, which every KV key node needs.

Cloudflare: List Page Rules

List a zone's Page Rules with their URL targets, actions, priority, and status. Page Rules is deprecated in favor of modern Rules but still served; Free zones get 3 rules.

Cloudflare: List R2 Buckets

List the account's R2 buckets with creation date, location, and storage class. Bucket MANAGEMENT only: reading and writing objects uses R2's S3-compatible API with separate R2 credentials and is not part of this connector.

Cloudflare: List Worker Scripts

List the account's deployed Worker scripts with names, handlers, usage model, and modification times.

Cloudflare: List Zone Rulesets

List a zone's rulesets (WAF custom rules, rate limiting, redirects, transforms) with id, kind, phase, and version. Read-only: this connector audits rulesets but does not write them.

Cloudflare: List Zone Settings

Read every setting on a zone in one call: SSL mode, security level, cache level, development mode, always_use_https, minify, and dozens more, each with its current value and whether the plan allows editing it.

Cloudflare: List Zones

List zones (domains) the token can see, with zone IDs, status, plan, and name servers. Filter by domain name, status, or account. The zone ID from here feeds every zone-level node (DNS, cache, settings, rules).

Cloudflare: Read KV Value

Read one key's stored value, returned verbatim as text at data.content (KV values are arbitrary bytes; JSON values arrive as a JSON string to parse downstream). A missing key returns 404. Values are capped at 200,000 characters in the node output.

Cloudflare: Verify API Token

Check the stored API token: returns its id, active/disabled/expired status, and validity window. Works with every scoped token regardless of permissions, so this is the credential health check to run first in any Cloudflare automation.

Action nodes 26

Cloudflare: Add List Items

Append items to a list - for a redirect list, the bulk redirect entries themselves. Asynchronous: the response returns an operation_id to poll with Cloudflare: Get Bulk Operation Status.

Cloudflare: Bulk Delete KV Pairs

DESTRUCTIVE: delete up to 10,000 keys from a namespace in one call. All listed values are gone permanently.

Cloudflare: Bulk Write KV Pairs

Write up to 10,000 key-value pairs in one call. Each item is an object with key and value plus optional expiration, expiration_ttl, metadata, and base64 fields. Existing keys are overwritten.

Cloudflare: Create DNS Record

Create a DNS record in a zone: A, AAAA, CNAME, MX, TXT, NS, and the other supported types. Reversible with Cloudflare: Delete DNS Record. Structured types like SRV and CAA take their data object via Extra body fields.

Cloudflare: Create KV Namespace

Create a Workers KV namespace. Reversible with Cloudflare: Delete KV Namespace.

Cloudflare: Create List

Create a list. kind=redirect makes a Bulk Redirect list (all plans: 5 lists and 10,000 URL redirects on Free); kind=ip makes an IP list usable in WAF expressions. Fill it with Cloudflare: Add List Items, then enable redirect lists via a Bulk Redirect rule in the dashboard.

Cloudflare: Create Page Rule

Create a Page Rule: a URL pattern plus the actions applied to matching requests (forwarding_url redirects, cache_level, always_use_https, and more). Counts against the plan's quota (3 on Free). Reversible with Cloudflare: Delete Page Rule.

Cloudflare: Create R2 Bucket

Create an R2 bucket (requires R2 to be enabled on the account). Reversible with Cloudflare: Delete R2 Bucket while the bucket is empty.

Cloudflare: Create Zone

Add a domain to a Cloudflare account. The response includes the two Cloudflare name servers to set at the registrar; the zone stays pending until they are. Reversible with Cloudflare: Delete Zone.

Cloudflare: Delete DNS Record

DESTRUCTIVE: delete a DNS record. The name stops resolving as caches expire and the record's exact contents are gone unless captured beforehand (Cloudflare: Get DNS Record or Export DNS Records).

Cloudflare: Delete KV Namespace

DESTRUCTIVE: delete a KV namespace and EVERY key-value pair in it, permanently. Workers bound to the namespace start failing their KV calls.

Cloudflare: Delete KV Value

DESTRUCTIVE: delete one key-value pair from a namespace. The stored value is gone unless read beforehand.

Cloudflare: Delete List

DESTRUCTIVE: delete a list and every item in it. Fails while rules still reference the list; remove those rules first. Redirects served from the list stop.

Cloudflare: Delete Page Rule

DESTRUCTIVE: delete a Page Rule permanently. Its redirect/cache behavior stops immediately and the rule configuration is gone unless read beforehand.

Cloudflare: Delete R2 Bucket

DESTRUCTIVE: delete an R2 bucket. The bucket must already be empty (R2 refuses to delete a bucket containing objects), but the bucket configuration and its name reservation are gone.

Cloudflare: Delete Worker Script

DESTRUCTIVE: delete a deployed Worker script permanently. Its routes stop being served and the code is gone from Cloudflare (keep the source in git). force=true also removes the script when other resources still reference it.

Cloudflare: Delete Zone

DESTRUCTIVE: remove a zone from Cloudflare. All DNS records, settings, and rules for the zone are deleted with it and cannot be restored; the domain stops resolving through Cloudflare immediately. Export DNS records first if there is any doubt.

Cloudflare: Edit Page Rule

Edit fields on an existing Page Rule; only the fields you fill are changed. Flipping status between active and disabled is the low-risk way to toggle a rule.

Cloudflare: Edit Worker Script Settings

Edit a Worker's metadata settings without redeploying code: toggle Logpush, set tags, wire tail consumers, or configure observability. Only fields you fill are changed.

Cloudflare: Edit Zone

Edit zone properties: pause or unpause Cloudflare on the zone (paused bypasses the proxy entirely), change the setup type, or set vanity name servers. Only fields you fill are sent.

Cloudflare: Edit Zone Setting

Change one zone setting, for example turn development_mode on, set security_level to under_attack during an incident, or set min_tls_version to 1.2. Reversible by writing the previous value back; read it first with Cloudflare: Get Zone Setting.

Cloudflare: Import DNS Records

Import DNS records into a zone from BIND zone-file text (the format Export DNS Records produces). Adds records; it does not delete existing ones. Returns how many records were parsed and added.

Cloudflare: Purge Cache

Purge Cloudflare's cache for a zone. Fill EXACTLY ONE of: purge_everything, files (URLs), tags, hosts, or prefixes - all purge modes are available on every plan. A purge cannot be undone, but the cache refills from the origin on the next requests; purging everything on a busy site briefly raises origin load. Deploy-webhook to purge is the classic automation.

Cloudflare: Rename KV Namespace

Rename a KV namespace. The namespace ID and all stored data are unchanged, so Workers bound by ID keep working.

Cloudflare: Update DNS Record

Update fields on an existing DNS record; only the fields you fill are changed. The classic dynamic-DNS move: point content at a new IP, or flip proxied on and off.

Cloudflare: Write KV Value

Write a value to a key (creating or overwriting it), with optional JSON metadata and expiry. KV is eventually consistent: edge reads can serve the old value for up to about 60 seconds. Values up to 25 MB; metadata up to 1024 bytes.

Connecting Cloudflare takes one credential

Cloudflare authenticates API requests with a scoped API token sent as a Bearer header. Tokens are created in the Cloudflare dashboard at dash.cloudflare.com under My Profile, then API Tokens (account-owned tokens live under Manage Account, then API Tokens), and each token carries the permission groups and zone or account scopes you choose at creation. API tokens are available on every Cloudflare plan, including Free. The legacy Global API Key is deliberately not supported by this connector: it is unscoped and all-powerful, and a scoped token is the right credential to hand an automation.

1

Sign in at dash.cloudflare.com, open My Profile, then API Tokens, and click Create Token. Start from a template like Edit zone DNS or build a custom token with the permission groups your workflows need, for example DNS edit, cache purge, zone settings edit, or Workers KV storage edit, scoped to the zones and account you want automated.

2

Copy the token when it is shown. Cloudflare displays the secret once at creation.

3

In Mattering Automations, open Connectors, click Add Connector, choose Cloudflare, and create a credential with the API token.

4

Click Test Connection. Mattering calls Cloudflare's token verification endpoint, which works for every scoped token regardless of its permission groups: an active result proves the credential, an error means the token is wrong, expired, or disabled.

5

All 56 Cloudflare nodes appear in the canvas palette. Run Cloudflare: List Zones first for zone IDs, and Cloudflare: List Accounts for the account ID that the KV, Workers, R2, and list nodes need.

Common uses for the Cloudflare MCP

Deploys that purge exactly the right cache

The classic Cloudflare automation: a deploy webhook fires and Cloudflare: Purge Cache clears what changed. The node takes exactly one mode per run, purge everything, specific URLs, tags, hostnames, or prefixes, and every one of those modes is available on every Cloudflare plan, so a Free zone can purge by prefix just like an Enterprise one. Purging by prefix or tag keeps the origin from absorbing a full cold-cache hit after every release.

An AI agent that manages DNS with a backup habit

Cloudflare: List DNS Records shows an agent what exists, Cloudflare: Export DNS Records captures the whole zone as a BIND file before anything changes, and then Cloudflare: Create DNS Record, Cloudflare: Update DNS Record, and a destructive-flagged Cloudflare: Delete DNS Record do the work. The update node is also the dynamic DNS move: point a record's content at a new IP or flip proxied on and off. If a change goes wrong, Cloudflare: Import DNS Records replays the exported records back in.

Incident response from a workflow

Cloudflare: Edit Zone Setting flips one setting at a time: set security_level to under_attack when traffic turns hostile, turn development_mode on while debugging cache behavior, or raise min_tls_version. Read the current value first with Cloudflare: Get Zone Setting so the workflow can write it back afterward, which makes the whole play reversible. Cloudflare: Edit Zone can pause Cloudflare on the zone entirely, which bypasses the proxy while you isolate a problem.

Workers KV as shared state for automations

Cloudflare: Write KV Value and Cloudflare: Read KV Value give workflows and Workers a common key-value store: an automation writes a release id or feature flag, and the Worker at the edge reads it on the next request. Cloudflare: Bulk Write KV Pairs loads up to 10,000 pairs in one call for lookup tables, and Cloudflare: List KV Keys walks a namespace by prefix. KV is eventually consistent, so edge reads can serve an old value for up to about 60 seconds after a write.

Bulk redirects at migration scale

Cloudflare: Create List with kind set to redirect makes the container, Cloudflare: Add List Items loads the source and target URL pairs, and because item writes are asynchronous, Cloudflare: Get Bulk Operation Status polls the operation to completed. The Free plan includes 5 redirect lists and 10,000 URL redirects, which covers most site migrations outright. One honest caveat: activating a list needs a Bulk Redirect rule enabled once in the Cloudflare dashboard, because this connector deliberately does not write rulesets.

Configuration audits that land in a database

Cloudflare: List Zone Settings reads every setting on a zone in one call, Cloudflare: List Zone Rulesets and Cloudflare: Get Zone Ruleset show the WAF custom rules with each rule's filter expression and enabled flag, and Cloudflare: List Page Rules covers the legacy layer. A scheduled sweep can write the results into a Mattering database and open a task when something drifted. The ruleset nodes are read-only on purpose: this connector audits the WAF, it does not edit it.

Example: A release goes live: back up DNS, repoint the record, purge what changed

Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.

1

Trigger: an inbound webhook from your CI pipeline fires when the deploy finishes, carrying the new origin address and the paths that changed.

2

Cloudflare: Export DNS Records captures the zone's full record set as a BIND file into the run, the backup that makes the rest of the workflow reversible.

3

Cloudflare: Update DNS Record points the record at the new origin. Only the fields you fill are changed.

4

Cloudflare: Purge Cache purges by prefix for the changed paths, so the origin serves fresh content without absorbing a full cold-cache hit.

5

Cloudflare: Write KV Value records the release id in a KV namespace the Worker at the edge reads on its next request.

6

The workflow opens a verification task in Mattering Productivity with the exported zone file attached, so a person confirms the site is healthy and has the rollback in hand if it is not.

Cloudflare MCP questions, answered

What Cloudflare operations are supported?

56 nodes on the Cloudflare API v4: 30 query nodes and 26 action nodes covering zones (list, get, create, edit, delete) and zone settings, DNS records with BIND export and import, cache purge in every mode, Workers KV (namespaces, keys, values, metadata, and 10,000-item bulk write and delete), deployed Worker scripts with script settings, R2 bucket management, read-only ruleset audits at zone and account level, Page Rules, and account lists with asynchronous item loading for Bulk Redirects. Deliberately not shipped: Worker code uploads, ruleset writes, R2 object reads and writes, and zone analytics.

Does this work with Claude, ChatGPT, or other AI assistants?

Yes. Mattering exposes these nodes over MCP, an open protocol for connecting AI assistants to tools, so any MCP-capable assistant, including Claude and ChatGPT, can connect to your workspace and use them. Matter, the assistant built into the platform, uses the same nodes natively.

Cloudflare publishes official MCP servers. Why use this connector?

Cloudflare does maintain its own MCP server family, and if all you want is an assistant talking to Cloudflare alone, it is a genuine option. This connector earns its place when Cloudflare is one step in a larger workflow: the same canvas holds your other connectors and the Mattering platform itself, so a deploy can purge cache and then open a verification task, a DNS change can wait behind a human approval step, and an audit can write into a database next to the rest of your work. Credentials live in one place with the rest of your connector secrets, and the same nodes serve visual automations and AI agents alike.

Do I need a paid Cloudflare plan?

Workers KV has a free tier, and R2 has one too, though enabling R2 means completing Cloudflare's R2 subscription checkout on the account first (free monthly usage is included, but it is a billing-enabled subscription rather than a toggle).

Can an automation purge the cache by tag, hostname, or prefix?

Yes, on every plan. Purge by URL, hostname, tag, prefix, and purge everything are all available across all Cloudflare plans, which is a real change: tag, hostname, and prefix purge were historically Enterprise-only, and a lot of older articles still say so. Throughput scales by plan, so on Free the hostname, tag, prefix, and purge-everything modes allow 5 requests per minute with up to 100 operations per request, while single-URL purges run far higher. Cloudflare: Purge Cache takes exactly one mode per run.

What are Cloudflare's API rate limits?

The global limit on the client API is 1,200 requests per 5 minutes per user, and it is cumulative across the dashboard and every token you own, so a heavy automation sweep and a person clicking around the dashboard draw from the same budget. Exceeding it blocks further calls for 5 minutes with HTTP 429. Nodes report the 429 on their output rather than silently retrying, so schedule wide sweeps apart and prefer one List call with filters over many Get calls.

Can an AI agent safely manage DNS records?

That is one of the connector's designed-for jobs, and the guardrails are real rather than implied. Cloudflare: Delete DNS Record is flagged destructive so workflows gate it, Cloudflare: Export DNS Records gives the agent a one-node backup of the whole zone before it changes anything, and Cloudflare: Import DNS Records replays that backup if a change goes wrong. The honest pattern is read, export, then write, with the delete step behind a condition or an approval task rather than a bare trigger.

What is deliberately not included?

Four things, each for a stated reason. Worker code uploads: deploying a multipart module bundle is not something to author safely on a canvas, so deploys stay in Wrangler or CI while the connector reads scripts, edits their settings, and can delete them. Ruleset writes: a mis-authored write replaces a zone's entire phase ruleset, which can drop the WAF in one call, so the ruleset nodes are read-only audits. R2 objects: reading and writing bucket contents uses R2's S3-compatible API with separate credentials, so this connector manages buckets only. Zone analytics: Cloudflare's REST analytics endpoints are sunset in favor of GraphQL, which is outside this connector's scope.

56 nodes waiting on the canvas

Put Cloudflare to work inside your workspace

Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.

All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. Cloudflare is a product of its respective owner; Mattering is an independent platform.

Start Free Explore our Products