Pay to create tests, not to run them.

• A test

  One flow on one browser profile. A checkout flow checked on desktop Chrome and on mobile Safari is two tests.

• Failed generations

  Don’t count. If we couldn’t write a working test, or found that the flow can’t run (a CAPTCHA, a missing 2FA secret), it takes no slot.

• A run

  One test executed once, whether you started it or a schedule did.

• An app

  One product you test, including up to 5 environments such as staging and production.

• Teammates

  Unlimited on every plan. We don’t charge per seat.

• Free

  One schedule, once a day. Your 5 tests daily use about 150 of your 500 monthly runs.

• Starter

  Your whole suite daily, plus a few critical flows hourly. 30 tests daily is about 900 runs a month; two of them hourly adds about 1,440, for about 2,340 of your 2,500.

• Team

  Schedules as often as every 15 minutes. One checkout flow every 15 minutes is about 2,880 of your 10,000 monthly runs.

• Enterprise

  Schedules as often as every 5 minutes, with tests and runs set in your contract.

• Over the limit

  Never billed and never blocked. Above 150% of your runs, they queue behind other runs until the month resets.

No. A passing run is plain Playwright and makes no AI calls, so every plan includes its runs. You pay for the tests you create. How a run works.

No. You don’t buy, top up or use up credits: a plan is a number of tests, runs a month and apps. A passing run makes no AI calls, so runs are included and going over is never billed. Writing tests and fixing broken ones uses AI, and both are in the plan price. Paid plans have a daily safety limit against runaway AI use: about 20 test generations a day on Starter and 60 on Team, shared with automatic fixes. Work over it waits for the next day (00:00 UTC) and is never billed. Need more? Email support@testwatch.dev. Ask for a higher limit.

Any time: start on Free, then upgrade from Settings → Billing in the app.

Nothing is blocked and nothing is billed. Between 100% and 150% of your runs you see a notice. Above 150%, your runs still happen, queued behind other customers’ runs until the month resets.

Runs from CI need Starter or higher: the GitHub Action, the CLI, the REST API, deploy hooks and preview-URL runs. Free runs on a daily schedule or by hand, and includes test inboxes, protected staging, suite export and JUnit download. Run TestWatch from CI.

A run of one test usually takes 5 to 20 seconds. Runs are queued, so a 30-test suite usually finishes in about 3 to 10 minutes, and longer at busy times. For a check on every pull request, tag your critical flows and run only those from CI, then run the whole suite on a schedule. Run only tagged tests from CI.

You can’t add a new test until you remove one or upgrade. Tests that failed to generate, or that we diagnosed as unable to run, don’t take a slot. What happens when a test breaks.

Each app can have up to 5 environments (for example staging and production), and its target host can change up to 5 times a day. That keeps an app meaning one product. How we handle your sites and data.

Nothing is deleted. If you are over the new plan’s limits, everything keeps running for 14 days; after that, schedules beyond the plan and the newest tests over the limit pause until you remove extras or upgrade.

Paid plans are for businesses and professionals. Prices are set in US dollars excluding VAT; Link, our payment partner, may charge the equivalent in your currency and adds VAT or sales tax where it applies. Buying for a company? Tick “buying as a business” at checkout and enter your VAT ID — Link applies the reverse charge where the rules allow it, and the invoice comes from Link.

Paid plans are for business and professional use: companies, sole traders and freelancers. Anyone can use the Free plan.

Cancel any time and keep the plan until the end of the period you paid for. Paid plans are business purchases, so unused time isn’t refunded, except for mistaken charges, a substantial outage, or when we end or change the service and you cancel. Refund policy.