Momentic

You install the CLI and point it at your application, locally or in CI. For scripted coverage you write a test as YAML in your repo: preset steps and assertions you control, with AI actions handling the parts that break when markup moves, such as finding an element by natural-language description rather than a brittle selector. Tests run against Chromium and Chrome, or against iOS simulators and Android emulators, on your infrastructure or on Momentic hosted machines. On every pull request your suite runs from GitHub Actions, GitLab or any other CI provider, and failures come back classified rather than as a wall of red. For exploratory work you use Mo instead: describe a flow in a prompt and the agent drives the app itself, returning findings with recordings and repro steps. An MCP server exposes the same capability to a coding agent, so a test can be written, run and debugged from the agent you are already in.

• •Mo, the autonomous QA agent - Explores a running app from a prompt and reports findings with screen recordings and reproduction steps. Carries a NEW badge on their own navigation, so it is the least proven half of the product
• •Agentic tests kept in your repository - AI actions get a goal, but preset steps and assertions live as YAML files you review in a pull request like any other code
• •Web, iOS and Android - Chromium and Chrome tests, plus iOS simulators and Android emulators, local or Momentic-hosted
• •Auto-healing and failure classification - Triage, failure recovery and result classification run as billed AI calls, each with a published per-call rate
• •CLI, CI and MCP server - Author and run tests from a terminal, wire the suite into any CI provider, or drive it from a coding agent over MCP
• •A published unit-rate card - Test steps, hosted browser minutes, emulator minutes and each class of AI call are all priced in the open next to the plans

Teams whose end-to-end suite has become the slowest and least trusted part of CI, and who would rather fix flakiness than delete the tests. Mobile teams get the most from the hosted emulators and simulators, and the five included phone numbers for SMS and OTP flows on the Scale plan matter if your signup path depends on them. The free plan is genuinely usable for a side project or a small suite, but its $50 monthly allowance is roughly four thousand test steps, so an active CI suite will exhaust it quickly and the realistic entry point is the $500 Scale plan. Usage billing also means the bill tracks the size of your suite rather than your headcount, which is the wrong direction if your suite is growing faster than your team. If you want a flat per-seat price you can forecast a year out, this is not that.