Scheduled API calls

Schedule API calls in plain English.

POST to your endpoint every weekday at 9:30 London time. Failed calls retry, every run lands in the history, and you hear about it when one keeps failing. No server, no redeploy, no UTC maths.

2.1k+ teams · Millions of schedules running right now

30 9 * * 1-5

Summarise to SlackPro

Perform a test to see what your endpoint returns. Example response.

Schedule your first call

Free for 2 schedules or monitors. No card.

Sound familiar?

Scheduled work shouldn't need its own server.

  • A server kept alive just to run one curl
  • GitHub Actions cron runs late when it's busy
  • UTC maths every time the clocks change
  • A failed call vanishes and nobody hears about it
  • Any method, headers and JSON body to any URL
  • Plain English or cron, in each schedule's own timezone
  • Automatic retries, and every attempt in the history
  • An email, or a Slack, Discord or Telegram message, when it keeps failing

Before and after it runs

See what every call returned, and hear when one fails.

Try the call once by hand, then let the schedule take over. Every run is logged, and a failing one comes to you instead of waiting to be found.

Test before you save

200 OK412 ms

Response

{
  "synced": 148,
  "ok": true
}

Test it before you save it

Perform test sends the real request from the form and shows the status, duration and response body, so the first scheduled run isn't the first try.

Failed· 14:44:15 · attempt 4 of 4

502 Bad Gateway (12,003 ms)

upstream api.your-app.com timed out

Alert sent to

Hear about failures first

When a schedule keeps failing (3 failed runs by default, you pick 1 to 100), Crontap emails you. On Pro it can also message Slack, Discord or Telegram.

Run history · last 14

FailedThu 05:00 · 50212,003 ms

13 ran · 1 failed

Every attempt on record

Status, code, duration, size and error text for each run and retry. 100 entries per schedule on Free, 1,000 on Pro.

Retries

A 503 at 09:30 is a retry at 09:31, not an outage.

Server errors, timeouts and dropped connections are retried automatically, about 1, 2 and 4 minutes apart. Pick what the endpoint does and watch the run play out.

nightly-sync · POST api.your-app.com/sync · daily 09:30

Run log

  1. 50309:30:001.2 s
  2. 20009:31:00388 ms

The retry a minute later worked. The run is logged as Retried, both attempts are in the history, and nobody is woken up.

Retries are on for every plan and stop before the next run is due. On Pro you choose 1 to 5 retries, a first wait of 30 seconds to 15 minutes, and whether 429s and other 4xx responses count.

The request

Send exactly what your endpoint expects.

  • GET, POST, PUT, PATCH, DELETE, HEAD or OPTIONS to any public URL.
  • Your own headers, so an API key in Authorization works like any other client.
  • A JSON body, edited with syntax highlighting.
  • Slow jobs are fine: a call can run up to an hour, as long as it ends before the next run is due.
  • Forward each result to a webhook of yours, on failure by default and on success if you want.

POST https://api.your-app.com/reports/daily

Headers

AuthorizationBearer ••••••••••••
Content-Typeapplication/json
X-Teamfinance

JSON body

{ "range": "24h", "notify": false }

daily 09:30 · Europe/London

Day

Crontap

Sent at UTC

UTC cron

Sat 27 Mar09:3009:30UTC+009:30
Sun 28 MarClocks go forward an hour09:3008:30UTC+110:30
Mon 29 Mar09:3008:30UTC+110:30

Crontap column: the schedule with its own timezone. UTC cron column: the same job pinned to UTC, as many cron hosts run it, shown in London time.

Timezones

09:30 in London stays 09:30 when the clocks change.

Every schedule has its own timezone, picked from the full IANA list. Crontap works out the UTC time for each run, so the London report and the Sydney sync can sit in one dashboard and both stay on local time.

AI integrations

Get the answer, not the JSON.

Add a prompt to a schedule and Crontap runs it over each response, then posts the result to a URL: a Slack webhook, Make, Zapier, n8n or your own endpoint. Text or JSON out, on success, and on failure if you ask.

Response from your schedule

GET api.your-app.com/stats/yesterday · daily 07:00

{  "date": "2026-11-15",  "orders": 312,  "revenue_eur": 18420,  "failed_payments": [    { "card": "visa", "country": "DE" },    { "card": "visa", "country": "DE" }  ],  "refunds": 0}

AI integration

Pro
Prompt
daily-numbersvia Slack webhook

312 orders yesterday, €18,420 in revenue, 2 failed payments, no refunds.

Example output

Pro, from $2.99/mo: one AI integration per schedule on daily or slower schedules. Ultra runs them hourly with no limit per schedule. Perform test runs it on any plan, so you can see the output before you upgrade.

Crontap MCP

Create and fix schedules from Claude or Cursor.

Ask for “POST to my sync endpoint every weekday at 9:30 London time” and your agent creates it, shows the next runs and reads the history when something fails. Same account, same limits.

Day to day

Run once, pause, group, copy.

  • Run once

    Fire a call at one set time, and delete the schedule after it runs if you want it gone.

  • Pause and resume

    From the list, the schedule page or your agent. A schedule that keeps failing for 7 days is paused for you, and you get an email.

  • Groups

    Name and colour groups, then drag schedules between them. Pro.

  • Duplicate

    Copy a schedule into a draft and change only what differs.

  • Search and sort

    Find a schedule by name or URL, sort by name, URL, created or updated.

  • Calendar view

    See when each schedule runs next on a calendar instead of reading cron.

Same tool, one bill

Everything else you’d juggle is already here.

Pricing

Start free. Upgrade when it matters.

Starter

$0forever

  • 2 schedules, monitors or heartbeats, hourly or slower
  • Automatic retries, 100 history entries per schedule
  • Email alerts
Save 33% yearly

Pro

$23.99/year$35.88

$2.00/mo billed yearly, or $2.99/mo monthly.

  • Every-minute schedules
  • Retries you tune, 1,000 history entries per schedule
  • Slack, Discord and Telegram alerts
  • AI integrations on daily schedules

No credit card to start. All plans

FAQ

Questions before you sign up.

Why not just use a cron on my server?
You can. Crontap is for when the server sleeps, scales to zero or isn't yours: it calls the URL from outside, retries, keeps the history and tells you when it fails.
What's the shortest interval?
Every minute on Pro, from $2.99/mo. The free plan runs schedules hourly or slower.
Do I have to write cron?
No. Describe it ("weekdays at 9:00", "first day of the month at 08:00") and Crontap turns it into cron, asking when something is unclear. If you prefer cron, paste a 5-field expression and you see what it means and the next 3 runs as you type.
What counts as a failure, and what gets retried?
A 4xx or 5xx response, a timeout or a connection error fails the attempt. Server errors, timeouts and connection errors are retried 3 times by default, about 1, 2 and 4 minutes apart. On Pro you can also retry 429s and other 4xx responses, and set 1 to 5 retries.
How do failure alerts work?
Crontap alerts you after a schedule fails 3 times by default; you can set anything from 1 to 100 and how often alerts are batched. Email works on every plan (at most once a day on Free). Slack, Discord and Telegram are on Pro. See alert channels and setup guides.
Can my AI agent manage schedules?
Yes, through Crontap MCP. Claude, Cursor, ChatGPT and other clients can create a schedule from plain English or cron, preview its next runs, pause or resume it and read its history. It uses your account and your plan's limits. Connect your agent.

2.1k+ teams · Millions of schedules running right now

Ready when you are

Your API, called on time.

Point Crontap at an endpoint, say when, and see it run. Retries, history and alerts come with it.

Free forever tier ・ No credit card required