Features · Monitor types

Every way a project can break. Watched.

A client's project is more than its homepage: an API behind it, a nightly job, a checkout flow and the providers it leans on. Probecast watches them from up to four regions — US West, US East, Europe and Asia — and files every monitor under the client it belongs to. On Indie, websites are checked from US East; Pro and Agency use all four.

At a glance

What each plan can watch.

  • Websites
    Availability by plan, in websites
    BasicIndieProAgency
    –1050100
  • APIs & endpoints
    Availability by plan, in endpoints
    BasicIndieProAgency
    –1030100
  • Ping, port & TCP
    • Basic not included
    • Indie included
    • Pro included
    • Agency included
  • Heartbeats
    Availability by plan, in heartbeats
    BasicIndieProAgency
    –1520
  • Browser flows
    Availability by plan, in browser flows
    BasicIndieProAgency
    –––10
  • Third-party services
    Availability by plan, in catalog services
    BasicIndieProAgency
    101060all

Numbers are monitors per account. Ping, port and TCP checks count toward the API allowance.

What you own

The parts of the project you shipped.

Each check records response time and status, alerts you when it fails and writes the incident into the client's history — which is what the monthly report reads later.

Websites

Uptime, certificate and content

Is the page up, how fast does it answer and is the SSL certificate about to expire? You choose how many days ahead the expiry warning arrives.

Content checks catch the broken deploy that still returns 200: the page must contain a phrase, must not contain another one — or both.

Pro and Agency check each site from all four regions; Indie checks from US East.

Availability by plan, in websites per account
BasicIndieProAgency
–1050100

APIs & endpoints

The endpoints your frontend calls

GET, POST or HEAD with your own headers, the status codes you expect and an optional keyword in the response. Paid plans choose which regions check each endpoint.

Built for health routes, webhook receivers and the API a mobile app depends on.

Availability by plan, in endpoints per account
BasicIndieProAgency
–1030100

Ping, port & TCP

Things that don't speak HTTP

A database port, a mail server, an SFTP drop or a game server. Ping answers “is the host there”; port and TCP answer “is the service listening”.

  • Basic not included
  • Indie included
  • Pro included
  • Agency included

Heartbeats

Cron jobs, workers and backups

Some failures make no noise: the nightly backup that stopped running, the queue worker that died quietly. Your job calls a ping URL when it finishes — if the call doesn't arrive within the expected window plus a grace period, you hear about it.

curl -fsS https://probecast.io/api/heartbeat/ping/<slug>
Availability by plan, in heartbeats per account
BasicIndieProAgency
–1520

Synthetic browser flows

Login, cart and checkout — as a user

A real browser walks through the flow that earns the client money and fails the check when a step breaks. Scripts run in an isolated sandbox; a failed run keeps a screenshot of the step that broke.

Describe the flow in plain words and AI drafts the script for you to review. Where we use AI →

Availability by plan, in browser flows per account
BasicIndieProAgency
–––10

Certificates & domains

Expiry dates nobody remembers

Certificate expiry comes with every website monitor. On every paid plan, Proactive monitors also watch certificates and domain registrations on their own — the renewal that lapses because the card on file expired at the client's registrar. Agency adds reminders for the renewals no check can see: contracts, licenses and security tests.

  • Basic not included
  • Indie included
  • Pro included
  • Agency included

What you depend on

When the problem isn't yours.

Half of the incidents a client notices start at a provider: the payment gateway, the email API, the auth service. Probecast already watches them for you.

Third-party services

1,000+ services in the catalog

Stripe, Shopify, OpenAI, Cloudflare, Vercel and the rest of the stack your projects lean on — checked continuously, with a public status page for each one. Add them to a client and their outages land next to yours, so you can tell the client which side broke.

See live incidents →

Availability by plan, in catalog services per account
BasicIndieProAgency
101060all

MCP servers

The tools your agents rely on

MCP servers are checked with a real protocol handshake, not a ping: a server that answers HTTP but can't initialize a session is never reported as up. They live in the same catalog and the same allowance.

MCP server status →

Availability by plan, in catalog services per account
BasicIndieProAgency
101060all

Set up from anywhere

From the dashboard, your terminal or your coding agent.

Websites, APIs and heartbeats can be created through the REST API, the CLI or the MCP server — the same quotas and the same client folders as the dashboard.

npx probecast add --type website --name "acme storefront" \
    --url https://acme.com --client "Acme"
Created website monitor 4f2a91c0-…

AI & developer tools →