Skip to main content
Every account needs a plan and a card before it can run a test. There is no free tier: a new account is taken to the pricing page first, and the first plan starts with a trial.
Prices and allowances on this page come from the product’s plan catalogue. In the app, Plans in the account menu opens the pricing page, where you choose or change a plan and buy credit packs. Settings → Billing and usage shows what your account is on.

​
Plans

Every plan also includes an AI execution allowance for each billing period, sized to its runs. It covers the model calls a run makes: planning, finding elements, and writing the report. The app shows it as a share used, never as a dollar amount. What counts as a run: every test that starts, passed, failed or needs review. A run that was cancelled before a browser ever started (for example, one that waited in line and was refused because every browser stayed busy) does not count. Every paid plan, and the trial, includes schedules, Slack and email alerts, the GitHub pull request integration, database assertions, mobile viewport and three-times replays of saved procedures, form fuzzing and the role boundary probe.

​
The trial

  • 3 days, on the plan you choose. Pick a plan on the pricing page and add a card. Nothing is charged until the trial ends.
  • Its own allowance: 10 runs, 30 browser minutes, and 1 run at a time, whichever plan you chose. The plan’s monthly allowance starts when the plan does.
  • Conversion: when the trial ends, the plan starts and the card is charged, unless you cancelled before then.
  • One trial per person. A second trial on a new account with the same email address is refused.
  • If you use the trial’s runs early, new runs are refused until the plan starts; the refusal names the day.
  • Credit packs are not sold during a trial.

​
Credit packs

One-time packs of runs and browser minutes, for a month that needs more than the plan includes.
  • Credits are used only after the month’s own runs or minutes run out.
  • They do not expire, but they can only be spent while a subscription is active.
  • They do not add to the AI execution allowance.
  • If your subscription is set to end, the pricing page asks you to confirm before selling you a pack.
There are no metered overages. When the allowance and credits are used up, new runs are refused until the period renews.

​
Usage and refusals

The usage meter in the app header shows runs, browser minutes and the AI execution allowance for the current period, and warns at 80%. Browser minutes count from the moment the browser starts to the moment the steps end. Planning, waiting in line for a browser, the report and the recording save do not count. Paid plans count usage over your billing period, not the calendar month.

​
Concurrency and capacity

Two limits decide whether a run starts now:
  1. Your plan’s concurrent runs (1, 2 or 5; 1 during a trial), counted across every workspace the plan pays for. A run over this limit is refused at once with the list of what is running. It is not queued.
  2. The service’s capacity: at most 5 runs hold a browser at once across every customer. When every test browser is busy, a run waits in line and the chat shows its place, an estimate when there is one, and a countdown. A run that is still waiting after 3 minutes at the front of the line, or 10 minutes in line altogether, does not start.
See Browser runs.

​
Seats and workspaces

Seats are counted per account: the distinct people across every team workspace the plan’s owner owns, plus open invitations to people who are not members yet. The owner counts once. Every account also has a personal workspace, which does not count against the shared-workspace limit.

​
Changing or ending a plan

  • Upgrades apply straight away, and the prorated difference is charged that day. Downgrades apply at the end of the billing period. The pricing page shows the new price and when it applies, and asks before it changes anything.
  • A failed payment leaves runs working for 72 hours, then pauses them until the payment succeeds.
  • When a subscription ends, new runs are refused. Projects, procedures and settings stay, and so does recent run history. Schedules are paused and start again by themselves when a subscription is active again.
  • History retention is enforced: runs older than the plan’s history period are deleted, with their screenshots, recordings and traces. An account without an active plan keeps 7 days of run history.