Scheduled HTTP requests to any public URL.

A cronjob for curl. Add a URL, a method and a schedule. The request goes out on time, a failure is retried, and a job that keeps failing pauses with one email. Free.

European by default. Requests leave from datacenters in the EU, and jobs and their history live in an EU-jurisdiction database. What we store

How it works

1 Add a URL and a schedule
A cron expression, a fixed interval, or a single moment. Headers and a body where the endpoint needs them.
2 The request goes out on time
From a short, fixed set of EU addresses, carrying a User-Agent that names the service and links to a stop page.
3 A failure arrives as one email
Three retries, then the job pauses and a single message lands with a resume link.

A job, as the dashboard shows it

One real row.

Status	Name	Method	Schedule	Next run	Last status
Active	Health check api.example.com/health	GET	every 5 minutes	in 4 min	200

Why the requests are safe to allow

Fixed EU addresses: A short, stable set of addresses, published at cronjob.dev/ips for allowlisting.
Requests that name themselves: Mozilla/5.0 (compatible; cronjob.dev/1.0; +https://cronjob.dev/stop/8f3k2m)
One click stops one job: The stop link in the User-Agent pauses the job behind the request. No account, no form, nothing to approve.
One record clears a whole zone: A TXT record at optout.<domain> with the value cronjob.dev keeps every job away from the domain and everything under it.
Responses stay unread: Status code, duration and timing are kept. The body is discarded.

Limits and timing

Jobs	50 per account. 5 during the first 24 hours.
Jobs per host	10. 2 during the first 24 hours.
Methods	GET, HEAD, POST, PUT, PATCH, DELETE, QUERY
Headers	Up to 8 per job
Body	16 KiB, POST, PUT, PATCH and QUERY only
Shortest interval	60 seconds. 5 minutes during the first 24 hours.
Timeout	10 seconds by default, up to 60
Addresses	IPv4. A host with no A record cannot be reached.
Timing	Within a few seconds of the scheduled time, not on the second.
Retries	Three retries after a failed run, then the job pauses and one email goes out.

Questions

What does a job cost?: Nothing, up to 50 jobs per account. No paid tier exists yet, so 50 is the ceiling rather than the start of an upsell.
Does the service read the response?: No. A run records the status code, the duration and how far the start drifted from the schedule. The body is discarded.
What counts as a failure?: A 4xx or 5xx status, a timeout, or a connection error. A 3xx is recorded and not followed.
Can a job reach a private address?: No. Private ranges and loopback are refused when the job is saved, and again before every run.
A server here is receiving requests. How do they stop?: The stop link in the User-Agent pauses that one job. A TXT record at optout.<domain> with the value cronjob.dev keeps every job away from the whole zone. Neither needs an account.
How long does history last?: 100 runs or 30 days per job, whichever comes first.
Is there an API?: Yes. Tokens live in settings, and the reference sits in the docs.