appstore-screenshot-mcp
Automates App Store screenshot capture from iOS Simulator, optimized for 2026 requirements with iPhone 6.9" and iPad 13" buckets.
README
appstore-screenshot-mcp
An MCP (Model Context Protocol) server for automating App Store screenshot capture from iOS Simulator. Optimized for the 2026 App Store requirements with 6.9" iPhone and 13" iPad buckets plus accepted fallback pixel sizes.
Features
- Simulator Management: List, boot, and shutdown iOS simulators with App Store-specific presets
- Screenshot Capture: Capture screenshots at native resolution with automatic status bar cleanup
- App Store Optimization: Pre-configured for iPhone 6.9" and iPad 13" buckets with all accepted pixel sizes
- Status Bar Control: Set clean status bar (9:41, full battery, WiFi) automatically
- App Management: Install, launch, and navigate apps for screenshot automation
- UI Interaction: Tap, swipe, scroll, and type to navigate apps to specific screens
- Video Recording: Record simulator interactions for App Store preview videos
- Appearance Control: Toggle dark/light mode, set locale, adjust text size
- Accessibility: Inspect UI hierarchy, find elements, configure accessibility settings
- Privacy Permissions: Grant, revoke, or reset app permissions (camera, location, photos, etc.)
- Push Notifications: Send simulated push notifications for UI testing
- Location Simulation: Set GPS coordinates or use preset locations
- System Alerts: Accept or dismiss permission dialogs automatically
- Keychain & Biometrics: Manage keychain, trigger Face ID/Touch ID
- Device Frames: Add device bezels for professional screenshots
- Validation: Verify screenshots meet App Store requirements
Installation
Via npx (recommended)
npx appstore-screenshot-mcp
From source
git clone https://github.com/your-org/mcp-screen-iOS.git
cd mcp-screen-iOS
npm install
npm run build
Requirements
- macOS with Xcode installed
- Xcode Command Line Tools (
xcode-select --install) - iOS 26+ Simulator runtimes
- Node.js 18+
Configuration
Claude Code
Add to ~/.claude.json or project .mcp.json:
{
"mcpServers": {
"appstore-screenshots": {
"type": "stdio",
"command": "npx",
"args": ["-y", "appstore-screenshot-mcp"]
}
}
}
Or via CLI:
claude mcp add appstore-screenshots npx appstore-screenshot-mcp
Codex CLI
Add to ~/.codex/config.toml:
[mcp_servers.appstore-screenshots]
command = "npx"
args = ["-y", "appstore-screenshot-mcp"]
Available Tools
Simulator Management
| Tool | Description |
|---|---|
list_simulators |
List available iOS simulators with filtering options |
boot_appstore_simulator |
Create and boot a simulator matching App Store size requirements |
shutdown_simulator |
Shutdown and optionally delete a simulator |
get_booted_simulator |
Get info about currently booted simulator(s) |
Screenshot Capture
| Tool | Description |
|---|---|
capture_screenshot |
Capture screenshot to file with status bar and notch options |
capture_screenshot_inline |
Capture and return as base64 image (for AI vision) |
capture_all_appstore |
Capture screenshots for all App Store sizes |
set_status_bar |
Configure status bar (time, battery, WiFi, cellular) |
clear_status_bar |
Reset status bar to system defaults |
validate_screenshot |
Verify screenshot meets App Store requirements |
App Management
| Tool | Description |
|---|---|
install_app |
Install a .app bundle on simulator |
launch_app |
Launch app by bundle ID |
terminate_app |
Stop a running app |
uninstall_app |
Remove app from simulator |
open_url |
Open URL or deep link (for navigation) |
list_apps |
List installed apps |
get_app_container |
Get app data/container path |
UI Interaction
| Tool | Description |
|---|---|
ui_tap |
Tap at specific coordinates (in points) |
ui_swipe |
Swipe from one point to another |
ui_scroll |
Scroll in a direction (up/down/left/right) |
ui_type |
Type text into focused input field |
ui_press_button |
Press hardware buttons (home, lock, volume, shake, toggleAppearance) |
get_ui_tree |
Get accessibility tree for finding element coordinates |
Video Recording
| Tool | Description |
|---|---|
start_recording |
Start recording video from the simulator |
stop_recording |
Stop recording and save the video file |
get_recording_status |
Check if recording is in progress |
Appearance Control
| Tool | Description |
|---|---|
set_appearance |
Set light or dark mode |
get_appearance |
Get current appearance mode |
toggle_appearance |
Toggle between light and dark mode |
set_locale |
Set locale and language (e.g., en_US, ja_JP) |
set_content_size |
Set Dynamic Type text size |
set_accessibility |
Configure accessibility options (reduce motion, bold text, etc.) |
Accessibility Inspection
| Tool | Description |
|---|---|
describe_ui |
Get the full UI accessibility hierarchy |
describe_point |
Get element info at specific coordinates |
find_element |
Search for elements by label, identifier, or type |
get_screen_info |
Get screen dimensions and device info |
App Store Screenshot Sizes (2026)
Privacy Permissions
| Tool | Description |
|---|---|
grant_permission |
Grant a permission (camera, photos, location, etc.) |
revoke_permission |
Revoke a previously granted permission |
reset_permission |
Reset permission to default (will prompt again) |
grant_all_permissions |
Grant all common permissions at once |
list_permissions |
List all available permission types |
Push Notifications
| Tool | Description |
|---|---|
send_notification |
Send a push notification with title, body, badge |
send_raw_notification |
Send a raw APNS JSON payload |
send_silent_notification |
Send a silent/background notification |
Location Simulation
| Tool | Description |
|---|---|
set_location |
Set GPS coordinates (latitude, longitude) |
set_preset_location |
Set to a preset city (Tokyo, London, NYC, etc.) |
clear_location |
Clear location override |
simulate_route |
Simulate movement along a GPX route |
list_preset_locations |
List available preset locations |
System Alerts
| Tool | Description |
|---|---|
accept_alert |
Accept/tap Allow on system dialogs |
dismiss_alert |
Dismiss/tap Don't Allow on dialogs |
trigger_siri |
Activate Siri |
send_memory_warning |
Send memory warning to apps |
trigger_icloud_sync |
Trigger iCloud sync |
Keychain & Biometrics
| Tool | Description |
|---|---|
reset_keychain |
Clear all stored credentials |
add_keychain_item |
Add a credential to keychain |
trigger_biometric |
Trigger Face ID/Touch ID match or failure |
enroll_biometric |
Enroll or unenroll biometrics |
Device Frames
| Tool | Description |
|---|---|
add_frame |
Add device bezel around screenshot |
add_background |
Add solid background with padding |
list_device_frames |
List available device frames |
App Store Screenshot Sizes (2024+)
Apple's 2026 requirements focus on two primary display buckets, with multiple accepted pixel sizes per bucket:
| Platform | Required Size | Accepted Dimensions (portrait) | Recommended Simulator |
|---|---|---|---|
| iPhone | 6.9" (or 6.5" if 6.9" isn't provided) | 1320x2868, 1290x2796, 1260x2736 | iPhone 17 Pro Max |
| iPad | 13" | 2064x2752, 2048x2732 | iPad Pro 13" (M5) |
The server validates against all accepted sizes for each bucket and reports whether a screenshot matches in portrait or landscape.
Example Workflows
Basic Screenshot Capture
1. Boot iPhone simulator: boot_appstore_simulator(deviceClass: "iphone_6.9")
2. Install your app: install_app(appPath: "/path/to/MyApp.app")
3. Launch app: launch_app(bundleId: "com.example.myapp")
4. Navigate to screen (if needed): open_url(url: "myapp://feature")
5. Capture screenshot: capture_screenshot(outputPath: "./screenshots/home.png")
Full App Store Screenshot Set
1. Boot both required simulators:
- boot_appstore_simulator(deviceClass: "iphone_6.9")
- boot_appstore_simulator(deviceClass: "ipad_13")
2. Install app on both (use UDIDs from step 1)
3. Navigate to desired screen
4. Capture all: capture_all_appstore(outputDir: "./screenshots")
Clean Status Bar Only
set_status_bar(
time: "9:41",
batteryLevel: 100,
batteryState: "charged",
wifiMode: "active",
wifiBars: 3
)
Navigate with UI Interactions
1. Launch app: launch_app(bundleId: "com.example.myapp")
2. Tap settings button: ui_tap(x: 350, y: 50)
3. Scroll down: ui_scroll(direction: "down", distance: 400)
4. Type in search: ui_type(text: "profile")
5. Toggle dark mode: ui_press_button(button: "toggleAppearance")
6. Capture screenshot: capture_screenshot(outputPath: "./screenshots/settings-dark.png")
Record App Store Preview Video
1. Boot simulator: boot_appstore_simulator(deviceClass: "iphone_6.9")
2. Launch app: launch_app(bundleId: "com.example.myapp")
3. Start recording: start_recording(outputPath: "./videos/preview.mp4")
4. Perform interactions: ui_tap, ui_scroll, etc.
5. Stop recording: stop_recording()
Capture Screenshots in Multiple Languages
1. Set locale: set_locale(locale: "ja_JP")
2. Terminate app: terminate_app(bundleId: "com.example.myapp")
3. Launch app: launch_app(bundleId: "com.example.myapp")
4. Capture: capture_screenshot(outputPath: "./screenshots/home_ja.png")
5. Repeat for other locales (fr_FR, de_DE, etc.)
Light and Dark Mode Screenshots
1. set_appearance(mode: "light")
2. capture_screenshot(outputPath: "./screenshots/home_light.png")
3. set_appearance(mode: "dark")
4. capture_screenshot(outputPath: "./screenshots/home_dark.png")
Grant Permissions Before Screenshots
1. grant_all_permissions(bundleId: "com.example.myapp")
2. launch_app(bundleId: "com.example.myapp")
3. -- App won't show permission dialogs, clean screenshots!
4. capture_screenshot(outputPath: "./screenshots/camera_feature.png")
Capture Notification UI
1. launch_app(bundleId: "com.example.myapp")
2. send_notification(bundleId: "com.example.myapp", title: "New Message", body: "You have a new message!")
3. -- Wait for notification banner
4. capture_screenshot(outputPath: "./screenshots/notification.png")
Location-Based Screenshots
1. set_preset_location(preset: "tokyo")
2. launch_app(bundleId: "com.example.myapp")
3. capture_screenshot(outputPath: "./screenshots/map_tokyo.png")
4. set_preset_location(preset: "new_york")
5. capture_screenshot(outputPath: "./screenshots/map_nyc.png")
Add Device Frame to Screenshot
1. capture_screenshot(outputPath: "./screenshots/home.png")
2. add_frame(inputPath: "./screenshots/home.png", shadow: true)
3. -- Creates home_framed.png with device bezel
Resources
The server exposes an MCP resource for device configurations:
appstore-screenshot-mcp://devices
This returns the supported App Store device presets with their identifiers and resolutions.
Development
# Install dependencies
npm install
# Build
npm run build
# Run in development (with watch)
npm run dev
# Type check
npm run typecheck
How It Works
The server wraps xcrun simctl commands and AppleScript to provide:
- Device Creation: Uses
simctl createwith exact device type identifiers for App Store sizes - Status Bar Override: Uses
simctl status_bar overridefor clean screenshots - Screenshot Capture: Uses
simctl io screenshotwith--mask=blackfor notch handling - App Control: Uses
simctl install/launch/terminatefor app lifecycle - UI Interaction: Uses AppleScript to control Simulator.app for taps, swipes, and keyboard input
- Video Recording: Uses
simctl io recordVideowith H.264/HEVC codecs - Appearance Control: Uses
simctl ui appearanceandsimctl spawn defaultsfor settings - Accessibility: Uses
simctl ui describefor UI hierarchy inspection - Privacy: Uses
simctl privacy grant/revoke/resetfor permission management - Notifications: Uses
simctl pushto send APNS payloads - Location: Uses
simctl location set/clearfor GPS simulation - Biometrics: Uses
notifyutilto trigger Face ID/Touch ID events - Device Frames: Uses ImageMagick (if available) or sips for image processing
All commands are executed via Node.js child processes with proper error handling.
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。