create-check
ChatGPTCreates a check of any supported type from a JSON configuration: API checks, browser and multi-step checks (with an inline script), heartbeat monitors, and URL, TCP, DNS, ICMP, traceroute, gRPC and SSL monitors. Call get-check-type-schema first to learn the check payload for the type. The check is created standalone, outside any Checkly CLI project, and starts running on its schedule when activated. Leave out optional settings the user did not ask for so their defaults apply; in particular, keep the SSL security baseline enabled unless the user explicitly asks to disable it.
create-check-group
ChatGPTCreates a check group. Put checks in it with create-check or update-check by setting groupId on the check. Group settings (locations, alerting, retries, API check defaults, environment variables) apply to every check in the group. Leave out settings the user did not ask for so their defaults apply.
create-status-page-incident
ChatGPTCreates a new incident on a status page, optionally notifying subscribers. This write is not idempotent; retrying the same call can create another incident and notify subscribers again.
delete-check
ChatGPTPermanently deletes a check of any type, including its configuration and result history. This cannot be undone. To stop a check from running while keeping it, use update-check with activated: false instead.
delete-check-group
ChatGPTPermanently deletes a check group AND every check in it, including their configuration and result history, like deleting a group in the Checkly web app. This cannot be undone. To keep the checks, first move them out of the group with update-check (groupId: null); to pause the group, use update-check-group with activated: false.
fetch-check-result-asset
ChatGPTFetches one selected check-result asset. Text assets may be returned inline with a byte limit; binary or large assets return a download link.
fetch-test-session-result-asset
ChatGPTFetches one selected test-session-result asset. Text assets may be returned inline with a byte limit; binary or large assets return a download link.
get-account-entitlements
ChatGPTReturns the resolved Checkly account plan and feature entitlement state.
get-account-environment-variable
ChatGPTGets one account-level Checkly environment variable by key. Secret values are always returned as null.
get-check
ChatGPTReturns the full configuration of one check of any type, in the Checkly public API v1 shape: schedule, locations, alert settings, and the type-specific configuration (request, script, heartbeat, ...). Use list-check-stats to find check ids.
get-check-group
ChatGPTReturns the full configuration of one check group, in the Checkly public API shape, and the checks in it. Group settings (locations, alerting, retries, API check defaults, environment variables) apply to every check in the group.
get-check-performance-summary
ChatGPTReturns a compact response-time performance summary for one check over a date range, including average, P50, P95, and P99 metrics when data is available.
get-check-result
ChatGPTReturns compact detail for a single check result. It includes useful run metadata but omits raw logs, traces, screenshots, videos, and artifact download URLs.
get-check-type-schema
ChatGPTReturns the JSON Schema of the check payload for one check type, as accepted by create-check (operation: create) or update-check (operation: update). Call this before creating or updating a check: every check type takes different configuration (e.g. request for API checks and uptime monitors, script for browser checks, heartbeat for heartbeat monitors).
get-error-group-root-cause-analyses
ChatGPTReturns a check or test-session error group with completed root cause analyses. Use this before triggering a new RCA.
get-root-cause-analysis
ChatGPTReturns a root cause analysis by RCA id. PENDING means polling should continue. COMPLETED is successful. FAILED means polling may stop for this attempt, but a late successful retry may supersede it.
get-status-page
ChatGPTGets one status page with its cards and services (v2) or components (v3) for incident targeting.
get-status-page-incident
ChatGPTGets one status page incident with services and incident updates.
get-test-session
ChatGPTReturns one test session and its compact result rows. Raw logs and asset URLs are omitted.
get-test-session-result
ChatGPTReturns compact detail for a single test-session result. It includes useful run metadata and small summaries, but omits raw logs and asset URLs.
get-usage-summary
ChatGPTReturns organization credits, pricing-rule breakdowns, and coverage gaps across one or more usage terms. Unknown prices and uncovered dates produce null credits. Optionally request a whole-term projection for an explicit usageTermsId; it covers all term accounts and pricing rules independently of summary filters, defaults to yesterday UTC, and reports failures separately. Requires usage:read permission.
get-usage-terms
ChatGPTLists organization usage contracts, including dates, budgets, pricing rules, and covered accounts. Use the returned ids to select terms or request whole-term projections. Requires usage:read permission.
invite-account-member
ChatGPTInvites a new user to the resolved Checkly account. Sensitive account administration action: requires checkly:account:invite plus OWNER or ADMIN account access, sends an invite email, and is not idempotent.
list-account-environment-variables
ChatGPTLists account-level Checkly environment variables and secrets. Secret values are always returned as null.
list-account-members
ChatGPTReturns current Checkly account members and pending invites. Supply search/filter arguments to look up members and invites using the public account-member list contract.
list-alert-channels
ChatGPTLists an account's alert channels (email, Slack, webhook, PagerDuty, Opsgenie, SMS, phone call) with the non-secret details that identify each one. Use the ids to set alertChannelSubscriptions when creating or updating a check with create-check or update-check.
list-check-groups
ChatGPTLists an account's check groups with their check counts. Use a group's id as groupId when creating or updating a check to put it in the group, and with get-check-group for its full configuration.
list-check-result-assets
ChatGPTReturns a normalized asset manifest for a check result. It does not fetch asset contents.
list-check-results
ChatGPTLists recent check results for one check using the same cursor ordering as the public v2 check results API. Returns compact rows only, without logs, traces, screenshots, or artifact downloads.
list-check-stats
ChatGPTReturns a paginated, filterable list of checks with their current pass/fail status. Set includeReliability to also fetch reliability analytics (availability, response-time percentiles, and ICMP latency/packet-loss) for each check over a quick range — this adds a ClickHouse query. Use this to triage status, and optionally reliability, across many checks at once instead of querying each individually.
list-status-page-incidents
ChatGPTLists status page incidents in the resolved account. Optionally filter to a status page and open/resolved/all incidents.
list-status-pages
ChatGPTLists status pages in the resolved Checkly account, including the cards and services (v2) or components (v3) needed for incident targeting.
list-test-session-result-assets
ChatGPTReturns a normalized asset manifest for a test-session result. It does not fetch asset contents.
list-test-session-results
ChatGPTLists compact test-session result rows with bounded pagination. Use this before get-test-session-result when a session contains many results. Raw logs and asset URLs are omitted.
list-test-sessions
ChatGPTLists recent test sessions for a Checkly account with the same filters and cursor ordering as the public v1 test sessions API. Each entry's errorGroupIds are the test-session error-group UUIDs — pass one to trigger-root-cause-analysis (type "test-session-error-group") or get-error-group-root-cause-analyses. The top-level running/passed/failed/cancelled arrays are testSessionResultIds that can be passed to get-test-session-result and the test-session asset tools. The sequenceIds object keeps the corresponding per-status sequence IDs for runner correlation and targeted cancellation.
list-usage-series
ChatGPTReturns cursor-paginated v2 usage by UTC period and optional account, usage term, pricing rule, and check type dimensions. Defaults to daily ungrouped rows. Include calculations, physical run counts, or observed monitor days when needed. Repeat the exact query with nextId. Requires usage:read permission.
resolve-status-page-incident
ChatGPTResolves a status page incident by posting a final resolved update. This write is not idempotent; retrying the same call can create another resolved update and notify subscribers again.
show-check-list
ChatGPTShow a paginated Checkly check-list component with current status. Use for “which checks are failing?” or “list my API checks”. Filters run on the server; at most 20 checks per page. No analytics, history or polling.
show-check-response-time
ChatGPTShow a response-time chart component for one check, with real time-bucketed average/P50/P95/P99 milliseconds. Use for “how did response time change?”. Defaults to the last 24 hours; an explicit startTime/endTime window is capped at seven days. One analytics query; no polling.
show-check-status
ChatGPTShow one check’s current status as an interactive Checkly check card, with its last-24-hour availability, P95 and P99. Use for “is this check healthy?”. One analytics query; no run history or polling.
trigger-checks
ChatGPTRuns existing (already-deployed) checks on Checkly ad hoc and records the run as a test session. Runs consume check quota and execute configured requests and code against external destinations, which may modify or delete data in the systems being checked. Select the checks with checkIds and/or tags — at least one selector is required. Set preview to true to dry-run the selection: it returns how many checks would run and a sample, without starting a session or consuming quota — recommended before a broad tag run on a large account. For OR-of-AND tag selection use tagGroups (an array of tag groups; a check matches if it has ALL tags in ANY one group), e.g. [["production","api"],["production","backend"]]. The simpler tags input is a single AND group. Returns the new test session; poll get-test-session with its id to follow progress, then use get-test-session-result and the asset tools for details. This only runs checks that already exist on the account. Only activated checks run: set allowDeactivated to true to include deactivated checks (and checks in a deactivated group); heartbeat monitors cannot be triggered. To create or deploy a new check, or to run local check code (browser, multistep, or Playwright suites), use the Checkly CLI.
trigger-root-cause-analysis
ChatGPTTriggers a new root cause analysis for a check or test-session error group and returns the pending RCA id. Consumes AI analysis quota. For eligible checks with automatic repair enabled, a completed analysis may start a repair that runs check code against external destinations and overwrites the check script after successful verification. Those runs may modify or delete data in the systems being checked. Call get-error-group-root-cause-analyses first and reuse an existing completed RCA when present.
update-check
ChatGPTUpdates a Checkly check. Either pass properties shared by every check at the top level (name, activated, muted, tags, useGlobalAlertSettings, runParallel, description, intent, aiAutoRepairEnabled), or pass check with a partial type-specific configuration (request, script, heartbeat, frequency, locations, alert settings, ...) whose schema get-check-type-schema returns with operation: update. Structured intent updates merge with existing intent, so omitted intent fields are preserved. Only send the settings the user asked to change; do not disable the SSL security baseline unless the user explicitly asks. This does not trigger a check run.
update-check-group
ChatGPTChanges the settings of a check group. Only the fields in group change; lists (tags, locations, alertChannelSubscriptions, environmentVariables) are replaced as a whole. Group settings (locations, alerting, retries, API check defaults, environment variables) apply to every check in the group.
update-status-page-incident
ChatGPTPosts a progress update to an existing status page incident. This write is not idempotent; retrying the same call can create another update and notify subscribers again.
whoami
ChatGPTReturns the authenticated Checkly user and account memberships visible to this MCP session.