Changelog
What we shipped, newest first. Follow it with the RSS feed.
Simpler pricing
Plans are now Free, Starter, Team and Enterprise. Yearly billing is $79 a month on Starter and $249 a month on Team. Starter schedules run as often as hourly and Team as often as every 15 minutes; Enterprise starts at $799 a month. There are no AI credits: each paid plan shows its daily safety limit as test generations a day (about 20 on Starter, 60 on Team).
Privacy and data controls
Each app can turn off screenshots and video. Download your data as JSON, and schedule your workspace or account for deletion yourself; a workspace can be restored for 7 days after you ask. AI calls go only to providers with zero-data-retention agreements with Vercel AI Gateway.
Stronger final checks
When a generated test ends on a weak check, generation makes one attempt to strengthen it. A test that already passes stays ready, with a note.
Chat-style generation
Describe a test and follow its generation in one chat thread. Regenerating a test shows its steps first, so you can edit them or say what changed before it starts.
A guided first test
Sign in with Google or GitHub and try TestWatch on a demo app first. Before generating, a quick HTTP check flags problems with your site; generation shows progress with an ETA and can email you when the test is ready.
Export your whole suite
Download all your tests as a runnable Playwright project, on every plan. It includes the config, the fixtures and an example env file with names only, no secrets.
Clean test data
Tests create their data with unique() and testEmail(), so every value is easy to find later. A signed cleanup webhook tells your app what each run created, on every plan.
Test email inboxes
Tests can receive email in a built-in inbox: verify links, magic links, password resets and emailed codes. On every plan.
Preview URLs and protected staging
Run tests against a preview deploy’s URL (Starter and up), and reach staging behind Basic Auth, custom headers, Vercel protection or Cloudflare Access on every plan. A fix proposed for a pull request can be accepted for that branch only.
Run from CI
Start runs from the GitHub Action, the CLI or the REST API and get one verdict and a JUnit report back. Signed deploy hooks start runs after a deploy, and signed webhooks send results out. Starter and up.
Full failure evidence
Runs keep console and network logs and a screenshot after each step. A test that passes on retry is marked flaky with its first attempt kept, a failure can be compared with the last pass, and you can cancel a run.
Failure emails on every plan
The workspace owner gets an email when a scheduled run fails, on every plan, until an alert channel takes over. Owners and admins also get a weekly summary email.
Tests run in a sandbox
Generated test code runs in its own sandbox for each job, with no network access except through our egress proxy.
Public pricing
The Free, Starter, Team and Business plans are on the pricing page. You pay for the tests you create; runs are included.
File uploads and downloads
Attach files to a test so it can upload them, and check the files your app lets people download.
Team roles and invites
Workspaces have Owner, Admin and Member roles, and you can invite teammates by email. Every account verifies its email before it can sign in.
Alerts and one Review inbox
Get failures in Slack or by email (Starter and up). The Review page collects every failure that needs a decision, from proposed fixes to likely bugs.