Write it in the browser
A full code editor with real Playwright completions, hover docs and type errors as you type. Press Save and run, and the result is on your screen moments later.
Scheduled Playwright monitoring
end2end.zip runs your Playwright tests in a real browser, on the schedule you set. Write them in the built-in editor or upload the ones you already have. When one fails you get the failing test, the error and a screenshot of what the browser saw.
Already have access? Sign in.
Checkout flow
3 passed 1 failed 37s total
expect(locator).toBeVisible() Locator: getByRole('button', { name: 'Pay now' }) Expected: visible Received: hidden
See it work
The real product, recorded start to finish. The run in it is a real one, against this site.
Why run tests on a schedule
The tests in your pipeline tell you a release was good on the day you shipped it. They say nothing about tonight. Most outages come from things that changed without a deploy:
Getting tests in
A test is an ordinary Playwright file either way. What you write here runs anywhere Playwright does, and what you already run in CI runs here.
A full code editor with real Playwright completions, hover docs and type errors as you type. Press Save and run, and the result is on your screen moments later.
Drop in the test files you already have. They are compiled the moment you upload, so a file that cannot run is rejected on the spot, with the line number, not at 3am.
Not a tester? Record your screen while you click through what must keep working, add a note, and assign it to a teammate to turn into a test. Nothing to install.
How it works
Write them in the editor or upload a zip. Set the site's base URL and any logins as encrypted project variables.
Every day at 07:00, every hour, every 5 minutes. Choose it in your own time zone, or trigger a run from CI after each deploy.
When a scheduled run fails you get an email, and every finished run can post a signed summary to your webhook.
What you get
A red mark tells you something broke. end2end.zip tells you what, where and what the page looked like, and makes sure the right person hears about it.
Each failed or timed-out test keeps its error message and a screenshot of the page at that moment.
Pick a local time and a cadence, down to every 5 minutes. Cron expressions are there if you want them.
A failure email for scheduled runs, plus a summary signed with HMAC-SHA256 posted to the webhook you choose.
Each project has its own trigger URL and secret. Call it after a deploy without handing your pipeline an admin key.
A failing test is retried automatically. One that only passes on a retry is marked flaky instead of quietly green.
Results appear test by test while the run is still going. Start a run by hand whenever you want an answer now.
Organisations, invitations and roles, with passwordless sign-in. Viewers can read every result and change nothing.
Every run gets its own sandboxed worker. Secrets are encrypted at rest, and a test can only reach the hosts you allow.
Questions
If a test file imports only @playwright/test and uses test, describe, the before/after hooks and expect, it runs as it is. The expect you get is Playwright's own, with the same auto-waiting matchers. Each file has to stand on its own: playwright.config.ts, custom fixtures and imports of your own helper modules are not supported.
Yes. Every project has an editor in the browser with real Playwright completions and type hints, so you can write a test, press Save and run, read the result and fix it without leaving the browser. You can also record your screen walking through a flow and assign the recording to a teammate to turn into a test.
In a headless Chromium browser on Cloudflare's Browser Rendering. Every run gets its own isolated worker, so your test code never shares a process with another customer's.
Yes. Store the credentials as project variables, which are encrypted at rest, and read them from process.env in your test. If a failing assertion quotes a secret back, it is redacted before the error is stored or sent anywhere.
Only the ones you allow. A test can load your project's base URL, its subdomains and any extra hosts you list. Everything else is blocked, including private network addresses, and redirects are checked at every hop.
A schedule can fire as often as every 5 minutes, in the time zone you pick. A test gets 60 seconds by default, which you can change per suite, and a whole run can take up to 13 minutes. You can also start a run on demand or from your CI with a project webhook.
30 days by default. You choose anything from 1 to 365 days per project, and runs and their screenshots are deleted once they pass that age.
end2end.zip is in early access and pricing has not been published yet. Join the waitlist and we will email you when access is available.
Put your most important flows on a schedule and get the evidence the moment one breaks.