faq
Questions, answered honestly.
What the agent does, what it never touches, and what lands on your verdict page.
Can my coding agent or CI run checks automatically?
Yes. Create an API key on your dashboard (Business plan) and POST to /api/checks with Authorization: Bearer cma_…. There's also an MCP server — claude mcp add checkmyapp — see mcp/README.md.
Do you look at my source code?
No. The agent uses your running app in a real browser like a first-time user; it never reads your repo.
What happens to the test-login credentials I provide?
Encrypted at rest, never logged, never shown in evidence. Used only so the agent can walk signed-in journeys.
Will it create real accounts or delete my data?
No. Signup is walked up to the final submit but not completed unless you allow it; destructive actions (delete, cancel) are off-limits unless you explicitly permit them.
What do I actually get?
A verdict page with journey-level findings, screenshots, network evidence, and downloadable Playwright specs. Daily Watch re-checks and files one ticket per new regression into your tracker.
How is the free tier limited?
A typical full first check costs us ~$0.50 of agent compute. Free = 1 check/day without signup, 3 total with a free account; paid plans lift the caps.
Which trackers are supported?
Linear today (GitHub next), on Growth and above.
Check your app → · first run is free, no signup.