Monitoring · checked 2026-10-19

Can I vibecode Checkly?

YES· one-shottable

Running API checks and Playwright scripts on a schedule, retrying before alerting, keeping the history and sending a message when something changes fits in one session: Playwright does the browser work and a few SQLite tables do the rest. What Checkly adds is where the checks run (6 to 22 locations outside your own network), alert routing into on-call tools and phones, status pages, and a team workflow with a CLI, Terraform and CI. For one person with a few sites, one location and email or ntfy alerts is enough, and the prompt includes a self-watch ping so a dead monitor doesn't stay silent.

$29/moprice

$348you'd save a year

one sittingbuild time

—people replaced it

Before you start

What you need on hand.

DifficultyIntermediate · Some system setup or keys.

TimeOne to two hours for the first version; Playwright on a VPS takes longer

Claude Code, Codex or CursorAny of them can run the prompt.
Node.js 22better-sqlite3 compiles a native module, so build tools must be installed.
Playwright's ChromiumInstalled with its system libraries; headless browsers use real memory, so size the server for them.
A machine at another providerA monitor on the same host as your site goes down with it.
A free Healthchecks.io checkAn outside heartbeat that tells you if the monitor itself stops.

Before you buildCheck Checkly Hobby first: 10 uptime monitors, 10,000 API runs and 1,000 browser runs a month cost nothing for one person. It stops, rather than bills, when the runs are used up.

1. Install one agent: Claude Code, Codex, Cursor.
2. Make an empty folder and open it in the agent.
3. Paste the prompt below, press enter, and answer its questions as it builds.

The prompt

Build me a self-hosted synthetic monitor for my own sites to replace Checkly. Requirements:

- Node.js 22 + TypeScript, one process: a scheduler plus Fastify serving Eta pages.
`npm start` listens on http://localhost:4880; /admin is behind one password from .env.
- Checks are files in ./checks: *.api.yaml (URL, method, headers, assertions on status,
JSON path, body text and latency) and *.spec.ts Playwright tests run with @playwright/test.
checks.yaml sets each check's interval (1 to 60 minutes), timeout and retries.
- Storage: better-sqlite3 in ./data/monitor.db: every run (time, ok, duration, error) kept
30 days, then rolled into daily percentages. After a failed Playwright run keep the
trace.zip and a screenshot in ./data/artifacts for 14 days.
- A check goes down after 2 failures in a row, the second 30 seconds after the first, and
recovers after 1 success. At most 2 browsers run at once; each run is cut off after 60 s.
- Alerts on state changes only: email (nodemailer), a Slack webhook and an ntfy.sh topic, all
optional in .env, plus one recovery message and at most one reminder per hour.
- Dashboard: each check with its status, a 24-hour strip of runs, a latency sparkline as
inline SVG, the last error and, for failed browser runs, the command that opens the trace
(`npx playwright show-trace <path>`).
- Self-watch: after every scheduler tick, ping HEALTHCHECK_URL (for example a free
Healthchecks.io check) so something outside this process tells me if it dies.
- No accounts, no telemetry, local-first: only my sites, the alert channels and
HEALTHCHECK_URL are contacted.
- Out of scope: multiple regions, private locations, status pages, SMS and phone alerts,
teams, deploying checks from a CLI.
- README: `npx playwright install --with-deps chromium`, how much memory one browser run
takes on my machine (measure it), and a warning to host it at a different provider from the
sites it watches, or one outage takes out both.

Opens your agent with the prompt filled in, you press enter. Or copy it and paste it into an empty folder.

Share on X

Max MiskowSponsoredAI agents, automations and MVPs, built for your company.Visit site ›

Then keep going

Next prompts, same folder.

The first prompt builds the core. Run these one at a time in the same folder to grow it into your own Checkly.

Add a second runner to the monitor in this folder. Requirements:

- A `runner` mode (`npm start -- --runner`) that has no UI: it pulls its check list from
  the main server over HTTPS with a bearer token and posts each result back.
- Each result carries the runner's name. A check only goes down when at least two runners,
  or the only one configured, report 2 failures in a row; show per-runner status in the UI.
- Tokens are stored hashed; a runner that hasn't reported for 5 minutes triggers an alert.
- Update the README with a systemd unit for the runner.

Add a public status page to the monitor in this folder. Requirements:

- /status with no login: one row per check marked public in checks.yaml, a 90-day bar of
  daily uptime and the incidents of the last 14 days.
- Incidents are created automatically from down periods, and I can add a note to one from
  /admin.
- No third-party scripts; the page works without JavaScript.
- Update the README.

Extend the Playwright checks of the monitor in this folder. Requirements:

- Per-check environment variables in checks.yaml, read from .env with a prefix per check,
  passed to the spec as process.env and never written to logs, traces or the database.
- Redact those values from error messages and from the stored trace before it is saved.
- An example spec that signs in, opens one page and asserts a known text.
- Update the README.

What you lose

Checks from 6 to 22 regions with round-robin or parallel scheduling
Alert routing into PagerDuty, Opsgenie, incident.io, Microsoft Teams, SMS and phone calls
Public status pages and dashboards tied to the monitors
The CLI, Terraform provider and AI-assisted root-cause analysis for a team
A year of aggregated history and private locations for internal services

Why people still pay

Checkly runs your Playwright tests as production monitors from outside your network, keeps the results, and pages the right people, all defined in the same repo as the code. A team that already writes Playwright tests gets monitoring without a new tool to learn, and nobody has to babysit a VPS full of browsers.

Moatinfrastructure scale integrations execution polish

Do the math

Checkly or your own copy?

Over

Your usage

Count my time3h ×$/h

Checkly, 1 yr$348

Your own copy, 1 yr$80running + build session

You keep$269pays for itself in the first month

Estimate · A laptop costs nothing but sleeps; a small always-on VPS with enough memory for headless Chromium runs roughly $5 to $12 a month. Healthchecks.io and ntfy.sh have free tiers. The build is token usage on your agent plan.

Leaving Checkly

Take your data, then cancel.

1 · Take your data out of Checkly

1Monitoring as code: in a CLI project logged in to your account, run npx checkly import. It generates code for the checks and other resources you managed in the web UI.
2Then npx checkly import apply links the generated code to your account, and npx checkly import commit finishes the import. Don't run deploy before that, or it duplicates your UI-managed resources.
3Playwright check suites are left out of an import because they are defined by the code bundle you uploaded; redeploy them from their source repository.
4Results: the CLI's checkly assets command lists and downloads logs, traces, videos and screenshots of runs.

2 · Cancel Checkly

1Open Account settings and the Billing page and switch the plan to Hobby; a third-party guide describes these steps and we didn't confirm them in Checkly's docs.
2The pricing page says a downgrade starts at the end of the billing cycle, so you keep the paid plan until then.
3To delete the account instead: Account settings, General, then Delete my account in the Danger zone. Only the account owner can, and a paid subscription must be cancelled first.

Deleting is permanent: all checks, data and account history go. Keep the generated code and any assets you need first.
Hobby is a hard cap, not a bill: when the monthly runs are used up, checks stop.
A downgrade to Hobby leaves one user and 10 uptime monitors, so remove the rest before you switch.
Keep the repository that holds your Playwright suites: the import can't turn an uploaded bundle back into source files.

Checked 2026-10-19 against Checkly's docs: 1, 2, 3, 4

The escape hatch

Don't feel like building it? These are already free.

Uptime KumaOpen sourceMIT self-hosted monitor for HTTP, TCP, ping and DNS with status pages and many notification channels; uptime checks, not Playwright flows.★ 93k · active Oct 2026 · docker to self-host
WebVisit ↗OpenStatusOpen sourceAGPL-3.0 status pages and uptime monitoring that you can host yourself or use hosted.★ 9.2k · active Oct 2026 · docker to self-host
WebVisit ↗GatusOpen sourceApache-2.0 developer-oriented health dashboard configured in YAML, with alerting and incident support.★ 12k · active Oct 2026 · docker to self-host
WebVisit ↗

All 4 free alternatives to Checkly ›

The numbers

Checkly pricing

Plan	Monthly	Annual, per month	What you get
Hobby	$0	$0	1 user, 10 uptime monitors at 2-minute intervals, 10,000 API check runs and 1,000 browser check runs a month, 6 locations, email, Slack and webhook alerts, 7 days of raw results.
Starter	$29	$24	3 users, 50 uptime monitors at 1-minute intervals, 25,000 API check runs and 3,000 browser check runs a month, 6 locations, SMS alerts, 7 days of raw results.
Team	$80	$64	10 users, 75 uptime monitors at 30-second intervals, 100,000 API check runs and 12,000 browser check runs a month, 22 locations, private locations, all alert channels, 30 days of raw results and a year of aggregates.
Enterprise	custom	—	Custom volumes, 1-second intervals, parallel scheduling, client certificates, 180 days of raw results and 25 months of aggregates.

Free tier: 10,000 API runs, 1,000 browser runs and 10 uptime monitors a month; one user; no private locations.
Billing: Monthly or annual (the page says annual saves up to 20%), prepaid upfront; overage is billed on the next invoice. Upgrades are prorated and take effect at once; downgrades start at the end of the billing cycle. Startups under two years old get 30% off and nonprofits 50%. Cards through Stripe.
Hidden costs: Runs are metered: a browser check every 5 minutes is about 8,640 runs a month, which is roughly $37 of overage on Starter beyond its 3,000. Extra uptime monitors cost $8 per 25. The Communicate add-on (alerting, status pages and dashboards) is $12 on Starter and $34 on Team when billed monthly, and Resolve (debugging and AI analysis) is $19 and $49.
Source: www.checklyhq.com

Also on the list

Better Stack$34/mo · UptimeYES

Cronitor$15/mo · Cron monitoringYES

Datadog$18/mo · MonitoringKINDA

Questions

Checkly, briefly

Not answered here? Tell us and we'll add it.

Ask a question ›

Hobby is free: 1 user, 10 uptime monitors, 10,000 API check runs and 1,000 browser check runs a month, from 6 locations.

Yes. An AI coding agent such as Claude Code, Codex or Cursor can build a usable personal version in one session with the prompt on this page, running on your own machine.

About $29 a month on the Starter plan, $348 a year (checked 2026-10-19).

Checks from 6 to 22 regions with round-robin or parallel scheduling; Alert routing into PagerDuty, Opsgenie, incident.io, Microsoft Teams, SMS and phone calls; Public status pages and dashboards tied to the monitors; The CLI, Terraform provider and AI-assisted root-cause analysis for a team; A year of aggregated history and private locations for internal services. If any of those matter to you, keep paying.