Skip to main content
Any coding agent — Claude Code, Cursor, OpenAI’s Codex — can work with noqa through the command line. It runs a command, reads the result, decides what to do next. Nothing to integrate, no dashboard in the way. The noqa-testing skill comes with it, so your agent already knows the commands and the workflows. An MCP server is on the way.

​
What your agent can do

  • Take a device — list iOS and Android devices, simulators and emulators, connect one, install a build.
  • See and act — read the current screen, then tap, swipe, drag or type by describing the target in plain language. See Device CLI.
  • Build the test suite — create and edit test cases, reusable flows, tags, and the app context your tests share.
  • Run the tests — hand saved cases to the noqa agent and read per-step results with screenshots.
Together that’s the whole job: your agent learns the app and writes the cases, our agent runs them. See Full-cycle testing.

​
Free and paid

The command surface depends on the account. Two tiers, each adding to the last (all commands require being signed in): The last row is a real split. The device and test-management commands your coding agent issues directly, deciding each next step itself. noqa runs create is different: it hands a saved case over to noqa, which then executes the whole scenario on the device autonomously.

​
Setup and reference

The CLI and the skill install from the app in a couple of clicks. Those steps, and every command with its options, live in the CLI reference.