Merge AI code on evidence, not hope

Two days waiting for QA can turn a small fix into another round of rebasing and context switching. See how a workflow using AI agents could cut manual QA effort by 30% across mobile, web, and TV, and learn how to validate the gains.

Appduct lets AI QA agents call functions registered by your team inside a running app. This article illustrates how an adjustable app clock and a simulated invitation event can prepare conditions for UI checks.

TesterArmy's e2e uses agent-device to drive mobile tests on iOS and Android. This article walks through goal-based actions, deterministic checks, the mobile engine, and replay caching, then shows how to initialize a test setup. Cached agent.act steps skip model calls; agent.assert continues to evaluate the current screen.

Mobile Dev brings live device previews, annotations, logs, and performance profiling into Codex. Explore the initial preview for iOS and Android development, with built-in agent-device support.

Jev is a model from TypeSafe that chooses between predefined options instead of generating text. We paired it with agent-device to explore what that approach could mean for AI-driven QA: reading app state, selecting actions, and checking the result. We walk through our proof of concept, how the two tools work together, and a recorded run that took 14 seconds and cost $0.0023 in model inference.

Code review and green CI checks do not prove a mobile app still works. Learn how cloud agents, agent-device, and runtime evidence can verify app behavior before merge.

Agent-device v0.20 turns known mobile journeys into scripts that agents can replay in seconds without further model calls. This article reports the experiments behind that workflow, based on 20 repeated runs, 38 Maestro flows, and 36 repair trials. It maps out when a task calls for settled observations, deterministic replay, live execution, repair, or a full re-recording as an app changes. It also covers how to validate repaired flows, prevent plausible but incorrect actions, and measure performance under realistic load.

Sniffler helps JavaScript and TypeScript teams run only the E2E tests affected by a change. It traces import and export relationships and reduces unnecessary CI work without asking teams to give up test coverage.

This article shows how we use agent-device with the Expensify app across three everyday mobile workflows: verifying bug fixes with recorded before/after evidence, measuring Sentry performance spans, and profiling React renders.

React Test Renderer’s deprecation forced React Native Testing Library to find a long-term replacement. In this article, Maciej Jastrzębski explains what it took to build a new JavaScript-only test renderer for RNTL and uses that process to show how React renderers actually work. It’s a practical look at React elements, Fiber, host instances, HostConfig, and React Reconciler for anyone who wants a clearer mental model of how JSX becomes UI.

Small edits to a SKILL.md can change agent behavior in ways that are hard to spot manually. This article shows how skillgym helps you test skills against real agent runners, validate the right behavior, and catch regressions before they reach production.

Agent React DevTools gives AI agents direct access to the React component tree, state, and performance data. Instead of relying only on UI snapshots, agents can now inspect what happens inside React and debug issues with proper context. This enables more reliable QA, better performance analysis, and deeper automation for React and React Native apps.

AI agents are powerful tools for mobile automation, but running them can get expensive and slow if they consume too many tokens. In this post, we break down how we optimized Agent Device to cut LLM token usage in half and speed up execution by 33%.

This article introduces agent-device, a new CLI for automating mobile UI workflows with AI agents. It explains what the tool is, how it helps developers with token-efficient snapshots and repeatable interactions, and why Callstack is building it to bridge exploratory agent behavior with deterministic replay across iOS and Android.

Introducing React Native Harness: a new way to test native modules directly on real devices and emulators. It brings together the speed and simplicity of Jest with the realism of on-device execution, giving developers a powerful, CI-ready tool to validate native behavior with ease. Harness makes it simple to ensure your native code works exactly as intended across every platform.

While many React Native testing tools like Jest and Appium are transferable to TV development, they lack built-in support for remote control interactions. For component tests, this means the standard library helpers are insufficient, requiring developers to build their own mocks that emit low-level hardware key events to simulate D-pad focus changes. For end-to-end tests, it requires distinct configurations and commands for native versus web-based TV platforms. This article explains how to create these custom JS test helpers from scratch and how to set up Appium to run automated E2E tests on emulators, real devices, and device farms.

Flow is a static type checker for JavaScript that can help catch errors in React code. This guide walks through practical examples of using Flow to type check React components, including stateless functions, class components, transferring props, and higher-order components.

Writing tests for UI can be tedious, but it's critical for ensuring code quality. This guide explores how AI tools like GitHub Copilot and ChatGPT can help make the process more efficient by suggesting test code for React components.

Detox is a new end-to-end testing library for React Native that allows you to simulate user interactions and make assertions on the app's interface. This guide walks through how to set up Detox with Jest and write your first tests to check for expected text and behavior.

Adding accessibility tests to your React Native app doesn't have to be a chore. This guide walks through setting up AccessibilitySnapshot, simplifying test creation, and automating the process with GitHub Actions.

Cypress is a new open-source testing tool that makes it fast and easy to write end-to-end tests for web apps. We walk through setting up Cypress to test core auth flows like login and logout in a real React app.

Snapshot testing can save you hours of headache when upgrading React Native apps. This step-by-step guide shows you how to set up snapshot tests for iOS, capture reference images, and validate UI changes at a glance.

Tired of waiting for tests to run on your local machine? This guide walks through how to set up cloud testing for React Native apps using XCUITest and AppCenter. You'll learn how to create test suites, submit your app, and verify the results across multiple devices.

Testing Redux reducers and actions in React Native doesn't have to be tedious. This guide shows how to use Jest snapshots to easily capture the expected output of your Redux code, simplifying the process of writing and maintaining tests.

Writing tests for React Native apps doesn't have to be painful. We explore how Jest's new snapshot testing feature makes it easy to catch bugs and verify your UI stays consistent as you iterate, without tedious manual comparisons.

Testing React Native apps on real devices can be challenging, especially for rare or older models. Cloud testing services like Microsoft's AppCenter allow you to easily run tests on hundreds of iOS and Android devices in parallel, without needing physical devices. This guide walks through how to set up a React Native project and configure AppCenter to run tests in the cloud using native frameworks like XCUITest.

Writing tests for web3 apps in React Native doesn't have to be daunting. This guide walks through configuring your project with Anvil, Wagmi, Viem, and Web3Modal, and shares practical tips for testing wallet connections, transactions, and edge cases without sacrificing speed or reliability.

Writing automated tests is key to shipping React Native apps with confidence. This guide shares practical tips for testing the most important parts of your app, from unit testing components to end-to-end flows, without getting bogged down trying to achieve 100% coverage.

Testing Redux with React Native Testing Library doesn't have to be complex. This guide walks through how to set up a custom render function to test your connected components in a real Redux environment, without relying on implementation details.

React Native Testing Library and Native Testing Library have officially merged into a single library under the Testing Library family. The new @testing-library/react-native package combines the best of both libraries while providing a smooth migration path for existing users.

React Native Testing Library 2.0 is out with new features to improve testing React Native components. It adds automatic cleanup, async act support, and scoped queries with the within operator. The release also removes some deprecated APIs and improves compatibility with community projects.

Writing accessible React Native apps for Android takes extra care. This guide shares practical tips to fix common issues like inline text links, icon sizes, text input placeholders, header titles, and WebView styling to ensure your app works well for all users.

Accessibility in mobile apps is about enabling equal access for people with disabilities. This guide shares practical tips for implementing accessibility in React Native, from design considerations like color contrast and font readability to testing with screen readers and other assistive tools.

As your React Native app grows, performance tends to degrade over time. This guide shows how to catch performance regressions early with Reassure, a new open-source tool for automating performance tests in CI/CD pipelines.

The React Native community has been using two separate testing libraries with near-identical interfaces. After much discussion, the libraries are finally merging into a single official package: @testing-library/react-native. The merge brings together the best of both worlds while making the migration process as smooth as possible.

React Native's core team recently introduced automated end-to-end (E2E) tests to improve the framework's reliability and catch regressions early. The effort was a collaboration between engineers from Callstack, Microsoft, and Meta, who worked together to research, propose, and implement a robust E2E testing setup using Appium, WebDriverIO, and Jest.

Writing unit tests for React components can be tricky, especially when checking implementation details instead of functionality. This article shows how to refactor tests using React Native Testing Library to make them more effective and focused on what users actually interact with.

Tired of performance regressions slowing down your React or React Native app? Reassure is an open source library that makes it easy to write performance tests, integrate them into your CI, and catch issues before they hit production.

Keeping React Native apps performant as they grow can be challenging. This guide walks through using the DMAIC process to consistently identify and fix performance issues, and catch regressions early with automated testing tools like Reassure.

Tired of flaky tests and complex mocking? Mock Service Worker (MSW) lets you intercept real network requests in your tests, making it easy to test components that rely on APIs. We share best practices and advanced techniques for using MSW effectively in larger projects.

Szymon Rybczak explains how TesterArmy tests web and mobile apps with AI agents, creates test plans for pull requests, and fits into QA workflows.

Zack Ebenfeld joins React Universe On Air to talk about how AI is changing React Native workflows, testing, QA, and day-to-day engineering work.

Find out how to monitor React and React Native performance better than ever before with Reassure's new features.

Explore Reassure, an Open Source tool for measuring performance regression in React Native apps.

Join Michał Pierzchała and Abi Thiao as they discuss the most interesting talks at App.js Conf 2022 and other React and React Native news.

The first episode of React Universe On Air (formerly known as The React Native Show) provides insights into React Native Testing Library. Mike and his guest give a quick glimpse into the very concept of Testing Library and tell the story behind React Native Testing Library.

This episode of the React Native Show dives into the recent introduction of End-to-End (E2E) testing in RNTester. Join us as we unpack the thought process, technical considerations, and practical implementation of this new approach to testing in React Native.

Michał Pierzchała shows how cloud-based QA agents can test real iOS and Android user flows and bring screenshots, recordings, and logs into pull requests.

AI can increase code output faster than teams can verify the real app. See why behavior review matters before merge.

Mike Grabowski will speak at React Summit 2026 about Agent Device, Callstack’s agent-native automation layer for giving AI coding agents a practical feedback loop across mobile, TV, and desktop apps.

Michał Pierzchała will speak at AI Coding Summit London 2026 about mobile QA agents for iOS and Android pull request workflows.

Join Callstack and Expo in San Francisco for practical talks on AI-native mobile development, Expo Agent, Cloud Agents and autonomous mobile QA.

Michał Pierzchała will present Automating Mobile QA with Cloud Agents at Ship London on 17 June 2026.

A discussion on Agent Device, AI-driven workflows, and testing mobile apps using simulators and accessibility data.

Join Szymon Chmal at WarsawJS #133 to explore cross-platform testing in React Native with a new approach powered by React Native Harness.

Learn how React Native Harness bridges the gap between mocking and E2E testing: real device tests, full native access, Jest-style simplicity.

This week’s Live Dev Session dives into React Native Harness with Szymon Chmal, explores the Next.js 16 Beta and Remix v3 “pre-drop,” and breaks down the biggest React and React Native updates from the community and React Alicante 2025.

Cross-platform testing in React Native made simple: a lightweight architecture to run and extend tests on any JavaScript-powered platform.

Developers often fix performance issues in React apps, but this is unsustainable and prone to regressions. Reassure, a testing library, helps spot problems before production, enhancing automated testing suites.

End-to-end tests are notoriously slow and costly, especially in mobile development. This talk introduces new tools and techniques to make E2E testing fast and efficient, so you can run tests on every pull request without draining CI resources.