> ## Documentation Index
> Fetch the complete documentation index at: https://docs.noqa.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> What agentic QA is, why vision-based testing is different, and who noqa is for.

Script-based tools like Appium or XCUITest break when the UI changes and can't interact with non-native elements like games, WebViews, or system dialogs. Manual testing works anywhere, but it doesn't scale.

noqa replaces scripts with an AI agent that interacts with your app the way a human tester would. It operates entirely from screenshots — no locators, no view hierarchies, no code. You describe the goal in plain English, and the agent decides what to tap, swipe, or type based on what it sees. This makes it compatible with any framework, and capable of testing things traditional automation can't — including games, canvas-based UIs, and third-party screens.

## Who noqa is for

* **Individual developers** — validate a feature before it goes to QA. Catch obvious regressions in minutes without waiting for a manual review cycle.
* **QA engineers** — automate repetitive regression flows and free up time for exploratory testing.
* **Mobile teams** — ship faster with automated coverage on every build. Run tests in CI/CD and get results before a release goes out.

## Your coding agent can do it for you

Claude Code, Cursor or Codex drive noqa from the command line: your agent explores the app on a device, writes the test cases from what it sees, and hands them to the noqa agent to run — you only say what to cover. See [Full-cycle testing](/guide/full-cycle-testing).

Ready to see how it works? [How noqa agent works →](/guide/how-noqa-agent-works)