WebView

Test hybrid apps that embed WebViews (login screens, payment flows, in-app browsers, Cordova/Ionic).

Prerequisites

• Android: The app must enable WebView debugging: WebView.setWebContentsDebuggingEnabled(true). React Native WebView provides a webviewDebuggingEnabled prop.
• iOS: The WKWebView must set isInspectable = true (required since iOS 16.4). React Native WebView provides a webviewDebuggingEnabled prop which sets this automatically. Tapsmith connects directly to the simulator’s WebKit Inspector — no external tools needed.

device.webview(packageName?: string): Promise<WebViewHandle>

Switch to a WebView context. Discovers available WebViews on the device and connects via CDP (Chrome DevTools Protocol).

const webview = await device.webview()
// or target a specific package when multiple WebViews are present
const webview = await device.webview("com.example.myapp")

Auto-waits for the WebView to appear up to the device timeout. Throws if no WebView is found.

device.native(): Promise<void>

Switch back to native context, closing the active WebView connection.

await device.native()
// Now you can interact with native elements again
await device.getByRole("button", { name: "Continue" }).tap()

WebViewHandle

Returned by device.webview(). All methods use CSS selectors and auto-wait for elements.

webview.click(selector: string): Promise<void>

Click an element in the WebView.

await webview.click("#login-button")
await webview.click(".submit-btn")

webview.fill(selector: string, value: string): Promise<void>

Fill an input element with text. Dispatches input and change events.

await webview.fill("#email", "[email protected]")
await webview.fill("#password", "secret123")

webview.textContent(selector: string): Promise<string>

Get the text content of an element.

const heading = await webview.textContent("h1")

webview.innerHTML(selector: string): Promise<string>

Get the inner HTML of an element.

webview.inputValue(selector: string): Promise<string>

Get the current value of an input element.

webview.getAttribute(selector: string, name: string): Promise<string | null>

Get an attribute value from an element.

webview.isVisible(selector: string): Promise<boolean>

Check if an element is visible (not display: none, visibility: hidden, or opacity: 0).

webview.evaluate<T>(expression: string): Promise<T>

Execute arbitrary JavaScript in the WebView and return the result.

const count = await webview.evaluate<number>("document.querySelectorAll('li').length")
const title = await webview.evaluate<string>("document.title")

webview.goto(url: string): Promise<void>

Navigate the WebView to a URL.

webview.title(): Promise<string>

Get the document title.

webview.url(): Promise<string>

Get the current URL.

webview.locator(cssSelector: string): WebViewLocator

Create a lazy locator for a CSS selector. Returns a WebViewLocator that can be used with expect() assertions.

const loginBtn = webview.locator("#login-button")
await loginBtn.click()
await expect(loginBtn).toBeVisible()

webview.close(): Promise<void>

Close the WebView connection. Usually called via device.native() instead.

WebViewLocator

Lazy reference to a CSS selector within a WebView. Supports actions and assertions.

Actions: click(), fill(value), textContent(), innerHTML(), inputValue(), getAttribute(name), isVisible()

WebView Assertions

expect(webview.locator(selector)) returns WebView-specific assertions:

await expect(webview.locator(".header")).toBeVisible()
await expect(webview.locator(".header")).toHaveText("Welcome")
await expect(webview.locator(".header")).toContainText("Welc")
await expect(webview.locator(".error")).toBeHidden()
await expect(webview.locator("#email")).toExist()
await expect(webview.locator("#email")).toHaveValue("[email protected]")
await expect(webview.locator("a")).toHaveAttribute("href", "/about")

All assertions support .not and a { timeout } option:

await expect(webview.locator(".spinner")).not.toBeVisible()
await expect(webview.locator(".loaded")).toBeVisible({ timeout: 10_000 })

MCP Server

Tapsmith includes a built-in MCP server that lets AI coding agents interact with devices, run tests, and inspect results. It supports two transport modes:

• SSE mode (via tapsmith test --ui) — agent shares the UI session with full test tree, results, watch mode, and mutual exclusion. 16 tools available.
• Stdio mode (client-launched tapsmith mcp-server) — standalone agent with its own headless test session, daemon, and device. Includes test discovery, results, watch mode, stop, and session info. 16 tools available.

See the MCP Server Guide for setup instructions, the full tool reference, and the recommended workflow.