Skip to main content
Glama

tickstem/mcp

Glama

MCP server for Tickstem — exposes cron job scheduling, uptime monitoring, heartbeat monitoring, and email verification as native tools for AI coding assistants (Claude, Cursor, Copilot, and any MCP-compatible agent).

Let your AI assistant register cron jobs, create uptime monitors with response assertions, manage heartbeat monitors, verify email addresses, and query results — while you write the app code.

Install

Download a pre-built binary (no Go required):

Platform

Binary

macOS (Apple Silicon)

tsk-mcp-darwin-arm64

macOS (Intel)

tsk-mcp-darwin-amd64

Linux (x86-64)

tsk-mcp-linux-amd64

Linux (ARM64)

tsk-mcp-linux-arm64

Windows (x86-64)

tsk-mcp-windows-amd64.exe

Download from Releases, make executable, and place on your PATH:

# macOS / Linux example
chmod +x tsk-mcp-darwin-arm64
mv tsk-mcp-darwin-arm64 /usr/local/bin/tsk-mcp

Or install with Go:

go install github.com/tickstem/mcp/cmd/tsk-mcp@latest

Related MCP server: @loggydev/mcp-server

Quick start

export TICKSTEM_API_KEY=tsk_your_key_here
tsk-mcp

The server speaks the Model Context Protocol over stdio — connect it to any MCP-compatible client.

Claude Code

Add to your ~/.claude/claude_desktop_config.json (or equivalent):

{
  "mcpServers": {
    "tickstem": {
      "command": "tsk-mcp",
      "env": {
        "TICKSTEM_API_KEY": "tsk_your_key_here"
      }
    }
  }
}

Available tools

Cron jobs

Tool

Description

list_jobs

List all cron jobs in the account

get_job

Get a cron job by ID

register_job

Register a new cron job (name, schedule, endpoint)

update_job

Update an existing job — only provided fields change

pause_job

Pause a job so it no longer fires

resume_job

Resume a paused or failing job

delete_job

Permanently delete a job and its execution history

list_executions

List execution history for a job, most recent first

Uptime monitoring

Tool

Description

list_monitors

List all monitors — status, URL, interval, SSL expiry, assertions

create_monitor

Create a monitor with optional response assertions (status code, response time, body)

get_monitor

Get a monitor by ID

pause_monitor

Pause a monitor so it stops polling

resume_monitor

Resume a paused monitor

delete_monitor

Permanently delete a monitor and its check history

list_monitor_checks

List recent checks — status, HTTP code, duration, error, SSL expiry

Response assertions

create_monitor accepts an assertions parameter — a JSON array of conditions that must all pass for a check to be considered up. When assertions are set they replace the default 2xx/3xx logic.

[
  { "source": "status_code",   "comparison": "eq",       "target": "200"              },
  { "source": "response_time", "comparison": "lt",       "target": "2000"             },
  { "source": "body",          "comparison": "contains", "target": "\"status\":\"ok\"" }
]

Source

Valid comparisons

Target

status_code

eq ne lt lte gt gte

integer string

response_time

eq ne lt lte gt gte

integer string (ms)

body

eq ne contains not_contains

plain string

Heartbeat monitoring

Tool

Description

list_heartbeats

List all heartbeats — status, token, interval, grace window, last ping time

create_heartbeat

Create a heartbeat monitor (dead-man's switch)

get_heartbeat

Get a heartbeat by ID

update_heartbeat

Update name, interval, or grace window

pause_heartbeat

Suppress alerts during planned downtime

resume_heartbeat

Resume alerting after a pause

delete_heartbeat

Permanently delete a heartbeat and its ping history

ping_heartbeat

Record a successful job run — no API key needed, token is the credential

list_heartbeat_pings

List recent pings for a heartbeat, most recent first

Email verification

Tool

Description

verify_email

Check syntax, MX records, disposable domain, and role-based prefix

list_verify_history

List past verification results for the account

Environment variables

Variable

Required

Description

TICKSTEM_API_KEY

Yes

API key from app.tickstem.dev

TICKSTEM_BASE_URL

No

Override API base URL (e.g. http://localhost:8080/v1 for local dev)

Local development

export TICKSTEM_API_KEY=tsk_your_key_here
export TICKSTEM_BASE_URL=http://localhost:8080/v1
go run ./cmd/tsk-mcp

SDKs

The MCP server handles infrastructure management from your AI assistant. For application code, use the SDK that matches your language:

Language

Install

Docs

Python

pip install tickstem

github.com/tickstem/python

Go (cron)

go get github.com/tickstem/cron

pkg.go.dev

Go (uptime)

go get github.com/tickstem/uptime

pkg.go.dev

Go (heartbeat)

go get github.com/tickstem/heartbeat

pkg.go.dev

Go (verify)

go get github.com/tickstem/verify

pkg.go.dev

Node.js

npm install @tickstem/cron

github.com/tickstem/node

All SDKs use the same API key.

Get an API key

app.tickstem.dev — free tier includes 1,000 cron executions, 5 uptime monitors, 5 heartbeat monitors, and 500 email verifications per month.

License

MIT

Available Tools

26 tools
create_heartbeatA
Destructive

Create a heartbeat monitor (dead-man's switch). Your job should POST to the returned ping URL after each successful run. If the ping stops arriving within the interval + grace window, Tickstem sends an alert.

Inputs: name (required), grace_secs, interval_secs. Returns: created_at, grace_secs, id, interval_secs, +6 more

create_monitorA
Destructive

Create an uptime monitor that polls an HTTP/HTTPS endpoint on a schedule and alerts on failure. Supports response assertions to validate that the response is correct — not just that the server responded. Assertions use source (status_code, response_time, body) + comparison (eq, ne, lt, lte, gt, gte, contains, not_contains) + target. When assertions are set they replace the default 2xx/3xx success logic. For HTTPS endpoints, SSL certificate expiry is captured automatically and an alert is sent 30 days before expiry. Pass assertions as a JSON string, e.g.: [{"source":"status_code","comparison":"eq","target":"200"},{"source":"body","comparison":"contains","target":""status":"ok""}]

Inputs: name (required), url (required), assertions, interval_secs, +1 more. Returns: assertions, created_at, id, interval_secs, +7 more

delete_heartbeatA
Destructive

Permanently delete a heartbeat monitor and all its ping history. This cannot be undone.

Inputs: heartbeat_id (required). Returns: message

delete_jobA
Destructive

Permanently delete a cron job and its execution history

Inputs: job_id (required). Returns: message

delete_monitorA
Destructive

Permanently delete an uptime monitor and all its check history. This cannot be undone.

Inputs: monitor_id (required). Returns: message

get_heartbeatB
Destructive

Get a single heartbeat monitor by ID. Returns its status, ping token, interval, grace window, and last pinged time.

Inputs: heartbeat_id (required). Returns: created_at, grace_secs, id, interval_secs, +6 more

get_jobB
Destructive

Get a cron job by ID

Inputs: job_id (required). Returns: created_at, description, endpoint, id, +6 more

get_monitorB
Destructive

Get a single uptime monitor by ID. Returns current status, SSL expiry date, assertion rules, and next scheduled check time.

Inputs: monitor_id (required). Returns: assertions, created_at, id, interval_secs, +7 more

list_executionsB
Destructive

List execution history for a cron job, most recent first

Inputs: job_id (required). Returns: executions

list_heartbeat_pingsB
Destructive

List recent pings for a heartbeat monitor, most recent first.

Inputs: heartbeat_id (required), limit. Returns: pings

list_heartbeatsB
Destructive

List all heartbeat monitors in the account. Each heartbeat has a token used for pinging and a status: active, paused, or failing.

Inputs: no named parameters. Returns: heartbeats

list_jobsA
Destructive

List all cron jobs in the account

Inputs: no named parameters. Returns: jobs

list_monitor_checksA
Destructive

List recent check results for an uptime monitor, most recent first. Each check includes: status (up/down/timeout), HTTP status code, response time in ms, error message, SSL certificate expiry date, and timestamp. Use this to diagnose failures or verify that a monitor is healthy.

Inputs: monitor_id (required), limit. Returns: checks

list_monitorsA
Destructive

List all uptime monitors in the account. Returns each monitor's status (active/paused/failing), URL, check interval, SSL expiry date, and assertion rules.

Inputs: no named parameters. Returns: monitors

list_verify_historyC
Destructive

List past email verification results for the account

Inputs: limit, offset. Returns: limit, offset, verifications

pause_heartbeatA
Destructive

Pause a heartbeat monitor. Alerts are suppressed while paused — useful during planned downtime or deployments.

Inputs: heartbeat_id (required). Returns: created_at, grace_secs, id, interval_secs, +6 more

pause_jobA
Destructive

Pause a cron job so it no longer fires

Inputs: job_id (required). Returns: created_at, description, endpoint, id, +6 more

pause_monitorA
Destructive

Pause an uptime monitor so it stops polling. No alerts will fire while paused. Use resume_monitor to restart.

Inputs: monitor_id (required). Returns: assertions, created_at, id, interval_secs, +7 more

ping_heartbeatB
Destructive

Ping a heartbeat to signal a successful job run. The token is the credential — no API key needed. Call this at the end of each successful execution.

Inputs: token (required). Returns: message

register_jobB
Destructive

Register a new cron job

Inputs: endpoint (required), name (required), schedule (required), description, +2 more. Returns: created_at, description, endpoint, id, +6 more

resume_heartbeatA
Destructive

Resume a paused heartbeat monitor. Alerting restarts immediately.

Inputs: heartbeat_id (required). Returns: created_at, grace_secs, id, interval_secs, +6 more

resume_jobA
Destructive

Resume a paused or failing cron job

Inputs: job_id (required). Returns: created_at, description, endpoint, id, +6 more

resume_monitorA
Destructive

Resume a paused uptime monitor. Polling and alerting restart immediately.

Inputs: monitor_id (required). Returns: assertions, created_at, id, interval_secs, +7 more

update_heartbeatA
Destructive

Update a heartbeat's name, interval, or grace window. Only provided fields are changed.

Inputs: heartbeat_id (required), grace_secs, interval_secs, name. Returns: created_at, grace_secs, id, interval_secs, +6 more

update_jobA
Destructive

Update an existing cron job. Only provided fields are changed

Inputs: job_id (required), description, endpoint, method, +3 more. Returns: created_at, description, endpoint, id, +6 more

verify_emailA
Destructive

Verify an email address: checks syntax, MX records, disposable domain list, and role-based prefixes

Inputs: email (required). Returns: created_at, disposable, email, id, +5 more

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 26 tools

    Disambiguation5/5

    Each tool is clearly namespaced by resource type (heartbeat, monitor, job, email) and action, so there is no real overlap between create_heartbeat, create_monitor, and register_job. The descriptions for similar lifecycle actions like pause_heartbeat, pause_monitor, and pause_job make their distinct targets obvious.

    Naming Consistency4/5

    The vast majority of tools follow a consistent verb_noun pattern with standard verbs like create, get, list, update, delete, pause, resume, and delete. Minor deviations include register_job instead of create_job and list_verify_history not matching the verify_email naming pattern.

    Tool Count4/5

    26 tools is on the high side, but the server covers four distinct domains: uptime monitors, heartbeats, cron jobs, and email verification. Each tool has a clear purpose, so the count feels slightly over the ideal range rather than bloated or redundant.

    Completeness3/5

    Heartbeat and cron job tooling is essentially complete, with full CRUD plus lifecycle operations. However, uptime monitors lack an update_monitor tool, so changing a monitor's URL, interval, or assertions requires deleting and recreating it, which loses history.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers