Menu
Mattering Automations connector

A StatusCake MCP with one node per API operation

36 nodes across uptime checks, SSL checks, pagespeed checks, heartbeat checks, maintenance windows, contact groups and the two monitoring location listers. That is every operation StatusCake API version 1.2.0 defines, with nothing merged and nothing dropped, so an agent can read outage history, open a planned maintenance window or fix a contact group without anyone opening the dashboard.

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

What the StatusCake MCP connector does

The StatusCake connector for Mattering Automations is 36 nodes on StatusCake API version 1.2.0, and the mapping to the API is one to one. Uptime checks get eight nodes: list, get, create, update, delete, plus the three series that hang off a check, its individual run history, its status periods and its alert history. SSL certificate checks, heartbeat checks, maintenance windows and contact groups get the five CRUD nodes each. Pagespeed checks get six, the five plus their own history series. Two more list the monitoring locations StatusCake runs uptime and pagespeed checks from. Eighteen nodes read and eighteen act.

Every capability is a node on the Automations canvas, and every node speaks MCP, the open protocol for connecting AI assistants to tools. Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. So an agent can read a week of outage periods with StatusCake: List Uptime Check Periods, work out which service was worst affected, and write the summary somewhere a person reads on Monday, with no dashboard session in the middle.

The three surfaces matter more here than on most connectors, because monitoring is the thing you want running when nobody is watching. The same nodes are steps on a scheduled workflow, so a nightly job can pause the right checks before a deploy and unpause them after. They chain with every other connector in one workflow, so an outage period can open a ticket, post to a channel and update a record without a bespoke script per destination. And the API token is held once in a shared workspace credential store rather than in each person's local assistant config, which matters for a token StatusCake documents no scopes for.

There is no first-party StatusCake MCP server to weigh this against. StatusCake's developer portal covers the API, official SDKs for Go, JavaScript, Python and Ruby, a Terraform provider and a cakectl command line tool it lists as coming soon, and mentions neither MCP nor the Model Context Protocol anywhere. Its GitHub organisation publishes those SDKs and the Terraform provider and nothing MCP shaped. So this connector is the practical way to put StatusCake behind an assistant today, and we would rather say that plainly than let the omission read as a claim.

Writes are form encoded, not JSON, and that is worth knowing even though the connector handles it. StatusCake's API description declares application/x-www-form-urlencoded at the top level, for every operation, and there is no JSON request body anywhere in the API. A body sent as JSON is never parsed into fields, so the API answers with a missing-required-parameter error that reads exactly like a bad node configuration. The connector sends every write the way StatusCake expects it, including the bracketed array spelling StatusCake's parameter guide documents, so this is a class of failure you get to skip rather than diagnose.

List fields replace rather than append, on every update, and this is the sharpest edge on the connector. Contact groups, tags, DNS IPs, regions, SSL alert days, email addresses, mobile numbers, integrations and the checks a maintenance window covers all behave that way. Adding one address to a contact group means reading the group first and sending the whole intended list back, because sending only the new one removes everybody else, silently, from a group whose entire job is telling people the site is down. An empty array is the documented way to clear a list, so the connector sends one when you mean it and leaves the field alone when you omit it.

Six nodes are flagged destructive and each says what is lost. Deleting an uptime, pagespeed, SSL or heartbeat check discards the check and everything recorded against it, and for an uptime check that is its entire run history, its status periods and its alert history, none of which any other endpoint can return once the id is gone. StatusCake documents no trash, no restore and no undo for any resource. Pausing is the reversible alternative and it keeps the history, which is what most workflows actually wanted. Pausing is not harmless either: a paused check is not run, so it cannot alert, and an unattended pause is an outage nobody is told about.

Two limits belong here rather than in the small print. StatusCake sells domain monitoring on its plan page and the API has no domain surface at all in version 1.2.0, no path and no definition, so this connector cannot create, read or alert on a domain monitor and no workflow should be designed as though it can. And this connector has no trigger nodes, so nothing here listens for a StatusCake alert. StatusCake's own alerting webhooks and integrations are a StatusCake feature configured in StatusCake; they are not a Mattering trigger, and the two are easy to blur. Workflows here run on a schedule, or on a trigger from another connector, and read.

One connection, three surfaces

  • Visual automations. Drag StatusCake 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 StatusCake node, ready to drop on the canvas

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

Read and query nodes 18

StatusCake: Get Contact Group

Get one contact group with its full recipient list: email addresses, mobile numbers, connected integration ids and the ping URL. Read the group with this before updating it, because every list on a contact group is replaced wholesale on update rather than appended to.

StatusCake: Get Heartbeat Check

Get one heartbeat check with its push URL, the period allowed between check-ins, when it was last pinged, its current status, contact groups and tags. This is how a workflow that has just created a heartbeat check retrieves the URL to hand to the job, since the create call returns only an id. The URL contains the check's push key, so pass it through a secret rather than logging or emailing it.

StatusCake: Get Maintenance Window

Get one maintenance window with its start and end times, the timezone those times are read in, how often it repeats, its current state and the exact list of uptime checks and tags it covers. Read the window with this before updating it, because the checks and tags lists are replaced wholesale on update.

StatusCake: Get Pagespeed Check

Get one pagespeed check with its configuration, the monitoring location it runs from and its latest recorded load time, page size and request count. Read a check with this before updating it so an update only changes what you meant to change.

StatusCake: Get SSL Check

Get one SSL check with the full certificate picture: issuer, valid from and valid until dates, cipher suite and scores, the flag block saying whether the certificate is broken, expired, revoked, missing or extended validation, and the list of mixed content resources found on the page. Use the flags object rather than parsing certificate_status text when branching a workflow.

StatusCake: Get Uptime Check

Get one uptime check in full: its type and target, check rate and timeout, the monitoring locations assigned to it, the string match and status codes it treats as failure, its contact groups and tags, when it last ran, where it runs next and its current status. Read a check with this before updating it so an update only changes what you meant to change.

StatusCake: List Contact Groups

List the contact groups on the account with the email addresses, mobile numbers, third-party integration ids and ping URL each one alerts. Every check type in this connector attaches its alerting by contact group id, so this is the node that populates those pickers and the place to look when a check is up and down as expected but nobody is being told.

StatusCake: List Heartbeat Checks

List the heartbeat checks on the account: the ones your own cron jobs and workers check IN to, rather than the ones StatusCake reaches out to. Each row carries the allowed period between check-ins, the current status, tags, contact groups and the push URL the job calls. Treat that URL as a secret: it embeds the check's push key, so anything that can call it can mark the job healthy.

StatusCake: List Maintenance Windows

List the maintenance windows on the account with their start and end times, timezone, repeat interval, current state and which uptime checks or tags they cover. Filter by state to answer the two useful questions separately: active tells you what is suppressing alerts right now, pending tells you what is scheduled.

StatusCake: List Pagespeed Check History

Every recorded run of one pagespeed check: when it ran, the load time in milliseconds, the page size in kilobytes, the request count, the simulated connection speed and a link to the saved HAR file for that run. StatusCake also computes the minimum, maximum and average of each measure ACROSS THE WHOLE REQUESTED WINDOW and returns them at data.metadata.aggregates, which is the number to quote for a period: recombining the per-run rows yourself gives a different and usually wrong answer, because these are per-run measurements rather than quantities that add up.

StatusCake: List Pagespeed Checks

List the pagespeed checks on the account with the latest recorded load time, page size and request count for each, plus whether the last run breached one of its alert thresholds and what the issue was. The latest_stats block is a snapshot of the most recent run only; use StatusCake: List Pagespeed Check History for a series.

StatusCake: List Pagespeed Monitoring Locations

List the servers StatusCake runs pagespeed checks from, with each one's region, region code, IP addresses and current status. Useful for allowlisting the addresses and for seeing which testing servers are healthy. Note these region codes are NOT the values the region field of a pagespeed check takes: that field uses its own eleven-code vocabulary (AU, CA, DE, FR, IN, JP, NL, SG, UK, US, USW), which is why the create and update nodes do not offer this list as a picker.

StatusCake: List SSL Checks

List the SSL certificate checks on the account with each certificate's issuer, validity dates, cipher, scores, mixed content findings and the alert thresholds configured for it. This is the node that answers what is about to expire: read data.data[].valid_until and compare it to now. Certificate scores and mixed-content detail are documented as paid-plan features, so those fields can be empty on a free account even though the check itself runs.

StatusCake: List Uptime Check Alerts

The alerts StatusCake actually sent for one uptime check, with the status that triggered each one, the HTTP status code seen and when it fired. Use it to answer whether anyone was told, which is a different question from whether the site was down: confirmation servers and the trigger rate delay mean a short outage can appear in the periods list with no alert here at all.

StatusCake: List Uptime Check History

Every individual run of one uptime check: when it ran, which monitoring server ran it, the status code that came back and the load time in milliseconds. Walk backwards through the series with before, or bound a period with before and after together, both as UNIX timestamps in seconds. Load time is a per-run measurement, so read the rows or chart them; adding them up means nothing and averaging them across days is only valid when each day contributed the same number of runs.

StatusCake: List Uptime Check Periods

The up and down periods of one uptime check: when each period started, when it ended and how long it lasted in milliseconds. This is the outage log rather than the raw run log, so it is the right source for a downtime report. Two things to know before doing arithmetic on it: a period that is still running has no end and no duration yet, and a period is filed under the moment it STARTED, so a window that cuts through an outage sees either all of it or none of it.

StatusCake: List Uptime Checks

List the uptime checks on the account with each one's current status, type, check rate, tags, contact groups and uptime percentage. This is the node the rest of the StatusCake connector leans on: uptime check pickers are populated from it and it is the cheapest way to confirm a credential works. Note that the uptime figure here is a percentage with no denominator in the payload, so it can be read and compared but never summed or averaged across checks.

StatusCake: List Uptime Monitoring Locations

List the servers StatusCake runs uptime checks from, with each one's region, region code, IPv4 and IPv6 addresses and whether it is currently up. Two uses: the region codes are the values the regions field of an uptime check takes, and the IP addresses are what you allowlist in a firewall so the checks are not blocked. Several servers share one region code, so the list is longer than the list of regions.

Action nodes 18

StatusCake: Create Contact Group

Create a contact group: the named bundle of email addresses, mobile numbers, integrations and a ping URL that checks alert through. Only the name is required, so a group can be created empty and filled later, but a check attached to an empty group alerts nobody. Returns only the new group's id at data.data.new_id, which is what you pass to the contact_groups field of any check.

StatusCake: Create Heartbeat Check

Create a heartbeat check: a check that goes down when your own job STOPS calling in, which is how you monitor a cron job, a backup or a queue worker that nothing outside can reach. Set period comfortably longer than the job's normal interval so a slow run does not alert. The response carries only the new id, so follow with StatusCake: Get Heartbeat Check to read the push URL the job needs to call.

StatusCake: Create Maintenance Window

Create a window during which alerting is suppressed for the uptime checks it covers, one-off or repeating daily, weekly, fortnightly or monthly. Two things decide whether it does what you meant. First, the times are read as LOCAL time in the timezone field and any UTC offset in the timestamps is ignored, so the timezone is not decoration. Second, at least one of tests or tags must be given or the window covers nothing at all. Only uptime checks can be covered: SSL, pagespeed and heartbeat checks keep alerting straight through a maintenance window.

StatusCake: Create Pagespeed Check

Create a pagespeed check that measures one page's load time, size and request count on a schedule from a chosen region. The three alert thresholds are all opt-in: a value of 0 means that alert is off, which is the default for all three, so a check created without them measures and never alerts. Returns only the new check's id at data.data.new_id.

StatusCake: Create SSL Check

Create an SSL certificate check for an https URL. The one field that catches people is alert_at: it must be a list of EXACTLY three numbers of days, such as [1, 7, 30], and any other length is rejected. SSL checks have no name in this API, so the URL is how they are identified everywhere. Returns only the new check's id at data.data.new_id.

StatusCake: Create Uptime Check

Create an uptime check of any of the seven supported types. The test_type field decides which of the other fields apply: HTTP and HEAD watch a URL, PING and TCP watch a host, DNS compares returned records against dns_ips, and SMTP and SSH check a service responds. Returns only the new check's id at data.data.new_id, so follow it with StatusCake: Get Uptime Check if the workflow needs the full record. How many checks an account may hold and how low the check rate may go are both capped by the StatusCake plan, so a rejected create is often a plan limit rather than a bad request.

StatusCake: Delete Contact Group

DESTRUCTIVE and irreversible. Deleting a contact group discards the group and its whole recipient list, and StatusCake documents no trash and no restore. The dangerous part is what happens to the checks: every uptime, SSL, pagespeed and heartbeat check that alerted through this group loses that route, and they keep running and keep going down while alerting nobody. Nothing warns you first. List the checks referencing the group id before deleting it, and reattach them to another group.

StatusCake: Delete Heartbeat Check

DESTRUCTIVE and irreversible. Deleting a heartbeat check discards the check, its recorded history and its push URL. Any job still calling that URL is now pinging a check that does not exist, and nothing tells you the job stopped: the monitoring goes quiet rather than going red. StatusCake documents no trash and no restore, and a recreated check gets a NEW push URL. If you only want the alerting to stop, use StatusCake: Update Heartbeat Check with paused set to true.

StatusCake: Delete Maintenance Window

DESTRUCTIVE and irreversible. Deleting a maintenance window removes the schedule and the list of checks and tags it covered, and StatusCake documents no trash and no restore, so a recurring window has to be rebuilt from scratch. The immediate effect matters as much as the loss: the checks it covered start alerting again, which during a planned outage means paging everyone. The checks themselves are untouched.

StatusCake: Delete Pagespeed Check

DESTRUCTIVE and irreversible. Deleting a pagespeed check discards the check and its entire recorded history: every load time, page size and request count ever measured for that page, and the saved HAR files behind them. StatusCake documents no trash and no restore. If you only want the measurements to stop, use StatusCake: Update Pagespeed Check with paused set to true.

StatusCake: Delete SSL Check

DESTRUCTIVE and irreversible. Deleting an SSL check discards the check and every certificate observation recorded against it, including the expiry history you would use to prove when a certificate changed. StatusCake documents no trash and no restore. If you only want the checks and the expiry warnings to stop, use StatusCake: Update SSL Check with paused set to true.

StatusCake: Delete Uptime Check

DESTRUCTIVE and irreversible. Deleting an uptime check takes its entire history with it: every recorded run and load time, every up and down period, and every alert ever sent for it. All of that is reachable only through this check id and no other endpoint can return it once the id is gone. StatusCake documents no trash and no restore. If you only want the checks to stop, use StatusCake: Update Uptime Check with paused set to true, which keeps the history.

StatusCake: Update Contact Group

Change an existing contact group: rename it or change who it alerts. Only the fields you fill are sent, but every list field REPLACES what is there rather than adding to it, so adding one email address means reading the group first with StatusCake: Get Contact Group and sending the whole list back with the new address included. Sending an empty array clears a list, which silently stops that channel for every check using the group. Succeeds with HTTP 204 and an empty body.

StatusCake: Update Heartbeat Check

Change an existing heartbeat check in place, including pausing it while a job is knowingly off and resuming it afterwards. Only the fields you fill are sent, and contact_groups and tags REPLACE the current lists rather than adding to them. Succeeds with HTTP 204 and an empty body.

StatusCake: Update Maintenance Window

Change an existing maintenance window: move it, change how often it repeats, or change which checks and tags it covers. Only the fields you fill are sent, but tests and tags REPLACE the current lists rather than adding to them, so read the window first with StatusCake: Get Maintenance Window and send the full intended list. Succeeds with HTTP 204 and an empty body.

StatusCake: Update Pagespeed Check

Change an existing pagespeed check in place, including pausing and resuming it and moving it to a different region. Only the fields you fill are sent, and contact_groups REPLACES the current list rather than adding to it. Changing the region changes what the numbers mean, so a history that spans the change is not a like-for-like series. Succeeds with HTTP 204 and an empty body.

StatusCake: Update SSL Check

Change an existing SSL check in place, including pausing and resuming it. Only the fields you fill are sent. alert_at and contact_groups REPLACE the current lists rather than adding to them, and alert_at must still be exactly three values when sent. Succeeds with HTTP 204 and an empty body, so read the check back with StatusCake: Get SSL Check if you need the result.

StatusCake: Update Uptime Check

Change an existing uptime check in place, including pausing and resuming it with the paused field. Only the fields you fill are sent, so everything else is left alone. Two things behave differently from the rest: the list fields (contact_groups, tags, dns_ips, regions) REPLACE what is there rather than adding to it, so send the full intended list and send an empty array to clear one; and the check's TYPE cannot be changed after creation, which is why test_type is not offered here. Succeeds with HTTP 204 and an empty body, so read the check back with StatusCake: Get Uptime Check if you need the result.

Connecting StatusCake takes one credential

One credential, one field. StatusCake's authentication guide documents a single scheme, an Authorization Bearer header carrying an API token, and you view and manage those tokens in the StatusCake account panel at app.statuscake.com/User.php. There is no account id, no key and secret pair, no region selector and no subdomain to supply. There is also nothing to scope: StatusCake's documentation describes no token scopes, no read-only tokens and no per-resource tokens, so a token is a full-account credential and anything holding it can reach every node on this page, including the six deletes. Treat it accordingly, keep the write and delete nodes behind conditions or an approval step in any agent-facing workflow, and rotate the token in the account panel if it is ever exposed. Mattering stores it encrypted, redacts it from workflow logs, and revoking it in StatusCake ends the connection immediately.

1

Sign in to StatusCake and open the account panel at app.statuscake.com/User.php, where API tokens are viewed and managed. Generate a token and copy it. StatusCake's guidance is that tokens must be kept private and that an exposed one should be replaced with a new one rather than reused.

2

Check the plan boundaries before you build, because they show up as failed creates rather than as blocked endpoints. StatusCake's pricing page lists the free tier as 10 uptime monitors, 5 minute test intervals, 1 page speed monitor, 1 domain monitor, 1 SSL monitor and alerts through integrations. How many checks you may hold and how low a check rate you may set are both plan questions, and the API rejects a rate the plan does not allow rather than the node doing it.

3

In Mattering Automations, open Connectors, click Add Connector and choose StatusCake. Create a new credential, paste the token into the single token field and save it.

4

Click Test Connection. Mattering calls the uptime check lister with a limit of one, which is the smallest read on the account. A 401 means the token is wrong or was revoked. A 429 means the account is over its rate limit rather than that anything is broken, so wait for the window and try again.

5

All 36 StatusCake nodes appear in the canvas palette. Run StatusCake: List Uptime Checks first: it needs no configuration, it is what every uptime check picker in the connector is populated from, and it confirms the credential in one call.

6

Note what a passing test does not prove. It is a read with no body, so it exercises none of the form encoding that every write depends on, and it touches only the uptime family. SSL, pagespeed, heartbeat, maintenance window and contact group nodes are separate paths, and a plan can limit any of them. Run the first write you care about deliberately rather than assuming it from a green connection.

Common uses for the StatusCake MCP

A morning reliability digest nobody has to assemble

StatusCake: List Uptime Checks returns every check with its current status, type, check rate, tags and uptime percentage in one call, which is the cheapest possible starting point. StatusCake: List Uptime Check Periods then reads the status transitions behind any check that looks bad, bounded to yesterday with the before and after parameters. An AI step turns that into plain language and a Mattering Productivity task or a chat message lands it where the team already looks. One caution belongs in the design: the uptime percentage is a rate with no denominator anywhere in the payload, no sample count and no window bounds, so it can be read and compared but never summed or averaged across checks. If you want a different window, request that window.

Suppressing alerts around a deploy, and turning them back on

StatusCake: Create Maintenance Window opens a suppression window over named uptime checks or over every check carrying a tag, which is the difference between a deploy that pages the on-call rota and one that does not. Two details decide whether it works. The start and end times ignore the UTC offset in the string you send: only the date and time parts are used and they are read as local time in the timezone field, so a window can run hours from where the timestamp appears to put it. And a window with neither checks nor tags is accepted by the API and suppresses nothing. StatusCake: Delete Maintenance Window is flagged destructive for a reason worth stating: the checks it covered start alerting again the moment it is gone.

Certificate expiry that reaches the person who can renew it

StatusCake: List SSL Checks returns each certificate's validity dates, issuer, cipher and configured alert thresholds, so a scheduled workflow can compare valid_until to today and escalate on its own terms rather than only through StatusCake's own alerting. StatusCake: Create SSL Check adds a new site to that coverage, and the field that fails most creates is alert_at, which must be exactly three numbers of days before expiry. Note the plan gate rather than debugging around it: StatusCake's pricing page lists mixed content alerts and certificate scores from its Superior tier, so the mixed content and score fields can be empty on a free account even though the check itself runs.

Watching the jobs that have no URL to check

A heartbeat check inverts the model: instead of StatusCake calling your service, your job calls StatusCake, and the check goes down when the call does not arrive inside the period. StatusCake: Create Heartbeat Check provisions one from a workflow, so a new nightly batch, sync worker or backup routine gets monitored the day it is written rather than the week after it silently stops. Handle the returned URL as a credential: it embeds the check's push key, so anything holding it can mark the job healthy. Route it through a secret rather than a log or a message. Deleting a heartbeat check is the quiet failure mode of this family, because the job keeps pinging a URL for a check that no longer exists and nothing reports that it stopped.

Keeping the on-call list correct without a dashboard trip

StatusCake: List Contact Groups and StatusCake: Get Contact Group read who gets told, and StatusCake: Update Contact Group changes it, so a rota change in an HR or scheduling system can reach monitoring the same day instead of at the next incident. Read the group before you write it: email addresses, mobile numbers and integrations each replace the stored list rather than adding to it. StatusCake: Delete Contact Group is the most under-rated destructive node on this connector, because it silently stops alerting for every check across all four families that used the group and the group id simply disappears from those checks. Nothing warns first.

Page speed trends read from the vendor's own arithmetic

StatusCake: List Pagespeed Checks gives the current picture per monitored page and StatusCake: List Pagespeed Check History walks the measurements over a window set with before and after as UNIX timestamps in seconds. The useful part is that StatusCake computes minimum, maximum and average over the window it was asked for and returns them alongside the rows, so a report reads the vendor's figure rather than reconstructing one from a page of samples. That distinction is deliberate everywhere in this connector: load times and response times are per-run measurements with no sample count to weight them by, so re-request the window you want rather than averaging averages.

Example: Suppress monitoring alerts for a release, then verify the site actually came back

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

1

Trigger: a schedule fires an hour before the release window, or a deploy pipeline calls a Mattering webhook. This connector has no trigger nodes, so nothing here waits on a StatusCake alert; StatusCake's own alert integrations stay configured in StatusCake.

2

StatusCake: List Uptime Checks returns the checks carrying the production tag, so the workflow acts on the current set rather than on a list somebody pasted in six months ago.

3

StatusCake: Create Maintenance Window opens a window over that tag, with the start and end times written in the timezone field the release is planned in, because the UTC offset in the timestamp is ignored and the timezone field is what decides when the window runs.

4

The release runs. A wait step or a callback from the pipeline holds the workflow until it is finished.

5

StatusCake: List Uptime Check History reads the most recent runs of the affected checks, returning the status code and load time of each individual run rather than a rolled up percentage, so the workflow can see whether the site answered correctly after the change.

6

A condition splits the outcome. If the recent runs look healthy the window is left to expire on its own schedule. If they do not, an AI step drafts what changed and a Mattering Productivity task is opened for the release owner, with StatusCake: List Uptime Check Alerts attached so the alert history is in the same place as the decision.

7

Nothing in this workflow deletes anything. Ending suppression early is StatusCake: Delete Maintenance Window, which is flagged destructive because the checks it covered start alerting again immediately, so it stays a deliberate step rather than a cleanup default.

StatusCake MCP questions, answered

What can this connector actually do in StatusCake?

36 nodes, 18 that read and 18 that act, one per operation in StatusCake API version 1.2.0. Reads cover listing and getting uptime checks, SSL checks, pagespeed checks, heartbeat checks, maintenance windows and contact groups, plus four series: an uptime check's individual runs, its status periods and its alert history, and a pagespeed check's measurement history, and the two listers for uptime and pagespeed monitoring locations. Writes cover creating, updating and deleting all six of those resource families. Not included, and worth knowing before you plan a workflow: domain monitoring, which StatusCake sells on its plan page and which has no API surface at all in version 1.2.0; server monitoring; public status pages and reporting; SMS credit management; team members, audit logs, account and billing; and alert integrations, which contact groups reference by id while no endpoint lists, creates or deletes one. There are also no trigger nodes.

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 work with StatusCake through them. Matter, the assistant built into the platform, uses the same nodes natively. The same nodes are also steps on the automation canvas, which is the part that matters for monitoring: the workflow an assistant helps you design then runs on a schedule with nobody present. Given that a StatusCake token has no scopes, the sensible split is to give an assistant the 18 read nodes and keep the writes, especially the six deletes, behind a condition or a human approval step.

Does StatusCake publish its own MCP server?

No. There is no first-party StatusCake MCP server. StatusCake's developer portal covers the API, official SDKs for Go, JavaScript, Python and Ruby, a Terraform provider and a cakectl command line tool listed as coming soon, and does not mention MCP or the Model Context Protocol anywhere; its GitHub organisation publishes those SDKs, the Terraform provider and a Telegraf agent repository, and nothing MCP shaped. So this page is not competing with a vendor server. What the connector adds over a hand-rolled script against the API is the platform around it: the nodes run on a schedule with nobody present, chain with every other connector in one workflow, hold the token once in a shared workspace credential store rather than in each person's local config, and carry destructive flags on the six nodes that need them.

What are StatusCake's rate limits?

StatusCake's rate limiting guide documents them plainly. Accounts without a subscription, or on a free plan, can make a maximum of 60 requests per minute. Accounts with a paid subscription have no such per-minute limit. All accounts, free or otherwise, are subject to a burst limit of 5 requests per second. Every response carries three headers: x-ratelimit-limit describing which limit applies, x-ratelimit-remaining, and x-ratelimit-reset in seconds until the window resets. Over quota, requests return HTTP 429 until the window resets. The burst limit is the one to design around even on a paid plan: a workflow that fans out and reads history for fifty checks at once will trip it. This connector does not retry or back off, so a 429 surfaces as an unsuccessful node with StatusCake's own response body, and the fix is to space the calls rather than to parallelise harder.

Which parts of StatusCake are gated by plan?

Capacity and features, rather than endpoints. There is no documented per-endpoint plan gate and no plan field in any response, so a plan limit shows up as a rejected create or an empty field rather than as a blocked node. StatusCake's pricing page lists the free tier as 10 uptime monitors, 5 minute test intervals, 1 page speed monitor, 1 domain monitor, 1 SSL monitor and alerts through integrations, so how many checks you can create and how low a check rate you can set are both plan questions. The API's uptime check rate accepts values down to 30 seconds and a continuous setting that lower tiers cannot use, and it is StatusCake that rejects a rate your plan does not allow, not the node. Mixed content alerts and certificate scores are listed from the Superior tier, so those fields on an SSL check can be empty on a free account even though the check runs. Custom locations are listed on the Business tier, so the regions field on an uptime check may be ignored or rejected on an account without it. Two families, heartbeat checks and maintenance windows, are not named on the pricing page in any tier, so we cannot tell you which plans include them and will not guess.

Can this connector monitor domains, servers or a status page?

No, and the domain one is the trap. StatusCake sells domain monitoring on its plan page, including one domain monitor on the free tier, and the API has no domain surface whatsoever in version 1.2.0: no path, no resource definition, and nothing for it in the vendor's own generated client. So domain expiry cannot be read, created or alerted on through this connector, and a workflow designed around it would be designed around something the API cannot do. The same is true of server monitoring, public status pages and reporting, SMS credit management, team members and audit logs. None of them appear in the API description, so none of them is a node here.

Can a workflow react to a StatusCake alert the moment a site goes down?

Not through this connector, and the distinction is worth being exact about. This connector has no trigger nodes: every node is a read or a write that a workflow calls, and nothing here listens. StatusCake does have its own alerting webhooks and integrations, but those are a StatusCake feature you configure inside StatusCake, pointed wherever you like, and they are not a Mattering trigger node. What works well today is the scheduled shape: a workflow that runs every few minutes and reads StatusCake: List Uptime Checks, or one that runs each morning over StatusCake: List Uptime Check Periods and StatusCake: List Uptime Check Alerts to summarise the night. The series nodes are built for that, taking before and after as UNIX timestamps in seconds so a run can read exactly the window since the last one.

What happens when a workflow deletes something in StatusCake?

It is gone, with no trash and no restore documented for any resource, which is why all six deletes carry a destructive flag and each names what is lost. Deleting an uptime, pagespeed, SSL or heartbeat check discards the check and everything recorded against it. For an uptime check that means its whole run history, its status periods and its alert history, all reachable only through that check id and none recoverable once it is gone. The reversible alternative is pausing the check with the matching update node, which stops the checks running and keeps the history intact, though a paused check cannot alert either, so an unattended pause is its own kind of outage. Deleting a maintenance window is not a data loss so much as an immediate behaviour change: the checks it covered start alerting again. Deleting a contact group silently stops alerting for every check across all four families that used it. And a heartbeat check is the quiet one: delete it and the job keeps pinging a URL for a check that no longer exists, with nothing reporting that the job stopped, while a recreated check gets a new push URL.

36 nodes waiting on the canvas

Put StatusCake 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. StatusCake is a product of its respective owner; Mattering is an independent platform.

Start Free Explore our Products