Add a new expert to the Remotion experts page
日本語の概要は準備中です。原文の説明を表示しています。
Generate Apple App Store screenshot pages as a Next.js app with html-to-image export at required resolutions. Screenshots are advertisements, not documentation — every screenshot sells one idea. Use when building App Store screenshots, generating exportable marketing screenshots for iOS apps, or creating programmatic screenshot generators. Triggers on 'app store screenshots', 'App Store', 'screenshot generator', 'iOS screenshots', 'marketing screenshots', 'phone mockup', 'ASO screenshots', 'app store assets', 'app listing', 'app store page', or 'app preview images'. Also use when someone is about to submit an app and needs store assets, or when they say 'make my app store page look good'. Includes iPhone and iPad mockup components with pre-measured dimensions.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Build a Next.js page that renders iOS App Store screenshots as advertisements (not UI showcases) and exports them via html-to-image at Apple's required resolutions. Screenshots are the single most important conversion asset on the App Store.
brand/creative-kit.md — Brand colors, fonts, logo pathsbrand/voice-profile.md — Copy tone and personalitybrand/positioning.md — Positioning angles for headline directionbrand/creative-kit.md for colors, fonts, and logo.brand/voice-profile.md for copy tone — this shapes headline style (playful vs. professional vs. minimal).brand/positioning.md for positioning angles — the primary angle informs the hero slide headline direction.Screenshots are advertisements, not documentation. Every screenshot sells one idea. If you're showing UI, you're doing it wrong — you're selling a feeling, an outcome, or killing a pain point.
If brand/ files provided answers (colors, fonts, voice tone), confirm them with the user and skip those questions. Only ask what's genuinely unknown. Do not proceed until you have all answers:
Based on the user's style direction, brand colors, and app aesthetic, decide:
IMPORTANT: If the user gives additional instructions at any point during the process, follow them. User instructions always override skill defaults.
Check what's available, use this priority: bun > pnpm > yarn > npm
# Check in order
which bun && echo "use bun" || which pnpm && echo "use pnpm" || which yarn && echo "use yarn" || echo "use npm"
# With bun:
bunx create-next-app@latest . --typescript --tailwind --app --src-dir --no-eslint --import-alias "@/*"
bun add html-to-image
# With pnpm:
pnpx create-next-app@latest . --typescript --tailwind --app --src-dir --no-eslint --import-alias "@/*"
pnpm add html-to-image
# With yarn:
yarn create next-app . --typescript --tailwind --app --src-dir --no-eslint --import-alias "@/*"
yarn add html-to-image
# With npm:
npx create-next-app@latest . --typescript --tailwind --app --src-dir --no-eslint --import-alias "@/*"
npm install html-to-image
The skill includes a pre-measured iPhone mockup at mockup.png (co-located with this SKILL.md). Copy it to the project's public/ directory. The mockup file is in the same directory as this skill file. No iPad mockup is needed — the iPad frame is CSS-only.
project/
├── public/
│ ├── mockup.png # iPhone frame (included with skill)
│ ├── app-icon.png # User's app icon
│ ├── screenshots/ # iPhone app screenshots
│ │ ├── home.png
│ │ ├── feature-1.png
│ │ └── ...
│ └── screenshots-ipad/ # iPad app screenshots (optional)
│ ├── home.png
│ ├── feature-1.png
│ └── ...
├── src/app/
│ ├── layout.tsx # Font setup
│ └── page.tsx # The screenshot generator (single file)
└── package.json
Note: No iPad mockup PNG is needed — the iPad frame is rendered with CSS (see iPad Mockup Component below).
The entire generator is a single page.tsx file. No routing, no extra layouts, no API routes.
// src/app/layout.tsx
import { YourFont } from "next/font/google"; // Use whatever font the user specified
const font = YourFont({ subsets: ["latin"] });
export default function Layout({ children }: { children: React.ReactNode }) {
return <html><body className={font.className}>{children}</body></html>;
}
Adapt this framework to the user's requested slide count. Not all slots are required — pick what fits:
| Slot | Purpose | Notes |
|---|---|---|
| #1 | Hero / Main Benefit | App icon + tagline + home screen. This is the ONLY one most people see. |
| #2 | Differentiator | What makes this app unique vs competitors |
| #3 | Ecosystem | Widgets, extensions, watch — beyond the main app. Skip if N/A. |
| #4+ | Core Features | One feature per slide, most important first |
| 2nd to last | Trust Signal | Identity/craft — "made for people who [X]" |
| Last | More Features | Pills listing extras + coming soon. Skip if few features. |
Rules:
Get all headlines approved before building layouts. Bad copy ruins good design.
<br />.| Type | What it does | Example |
|---|---|---|
| Paint a moment | You picture yourself doing it | "Check your coffee without opening the app." |
| State an outcome | What your life looks like after | "A home for every coffee you buy." |
| Kill a pain | Name a problem and destroy it | "Never waste a great bag of coffee." |
page.tsx
├── Constants (IPHONE_W/H, IPAD_W/H, SIZES, design tokens)
├── Phone component (mockup PNG with screen overlay)
├── IPad component (CSS-only frame with screen overlay)
├── Caption component (label + headline, accepts canvasW for scaling)
├── Decorative components (blobs, glows, shapes — based on style direction)
├── iPhoneSlide1..N components (one per slide)
├── iPadSlide1..N components (same designs, adjusted for iPad proportions)
├── IPHONE_SCREENSHOTS / IPAD_SCREENSHOTS arrays (registries)
├── ScreenshotPreview (ResizeObserver scaling + hover export)
└── ScreenshotsPage (grid + device toggle + size dropdown + export logic)
const IPHONE_SIZES = [
{ label: '6.9"', w: 1320, h: 2868 },
{ label: '6.5"', w: 1284, h: 2778 },
{ label: '6.3"', w: 1206, h: 2622 },
{ label: '6.1"', w: 1125, h: 2436 },
] as const;
Design at the LARGEST size (1320x2868) and scale down for export.
If the user provides iPad screenshots, also generate iPad App Store screenshots:
const IPAD_SIZES = [
{ label: '13" iPad', w: 2064, h: 2752 },
{ label: '12.9" iPad Pro', w: 2048, h: 2732 },
] as const;
Design iPad slides at 2064x2752 and scale down. iPad screenshots are optional but recommended — they're required for iPad-only apps and improve listing quality for universal apps.
When supporting both devices, add a toggle (iPhone / iPad) in the toolbar next to the size dropdown. The size dropdown should switch between iPhone and iPad sizes based on the selected device. Support a ?device=ipad URL parameter for headless/automated capture workflows.
Each screenshot is designed at full resolution (1320x2868px). Two copies exist:
transform: scale() via ResizeObserver to fit a grid cardposition: absolute; left: -9999px at true resolutionThe included mockup.png has these pre-measured values:
const MK_W = 1022; // mockup image width
const MK_H = 2082; // mockup image height
const SC_L = (52 / MK_W) * 100; // screen left offset %
const SC_T = (46 / MK_H) * 100; // screen top offset %
const SC_W = (918 / MK_W) * 100; // screen width %
const SC_H = (1990 / MK_H) * 100; // screen height %
const SC_RX = (126 / 918) * 100; // border-radius x %
const SC_RY = (126 / 1990) * 100; // border-radius y %
function Phone({ src, alt, style, className = "" }: {
src: string; alt: string; style?: React.CSSProperties; className?: string;
}) {
return (
<div className={`relative ${className}`}
style={{ aspectRatio: `${MK_W}/${MK_H}`, ...style }}>
<img src="/mockup.png" alt=""
className="block w-full h-full" draggable={false} />
<div className="absolute z-10 overflow-hidden"
style={{
left: `${SC_L}%`, top: `${SC_T}%`,
width: `${SC_W}%`, height: `${SC_H}%`,
borderRadius: `${SC_RX}% / ${SC_RY}%`,
}}>
<img src={src} alt={alt}
className="block w-full h-full object-cover object-top"
draggable={false} />
</div>
</div>
);
}
Unlike the iPhone mockup which uses a pre-measured PNG frame, the iPad uses a CSS-only frame. This avoids needing a separate mockup asset and looks clean at any resolution.
Critical dimension: The frame aspect ratio must be 770/1000 so the inner screen area (92% width × 94.4% height) matches the 3:4 aspect ratio of iPad screenshots. Using incorrect proportions causes black bars or stretched screenshots.
function IPad({ src, alt, style, className = "" }: {
src: string; alt: string; style?: React.CSSProperties; className?: string;
}) {
return (
<div className={`relative ${className}`}
style={{ aspectRatio: "770/1000", ...style }}>
<div style={{
width: "100%", height: "100%", borderRadius: "5% / 3.6%",
background: "linear-gradient(180deg, #2C2C2E 0%, #1C1C1E 100%)",
position: "relative", overflow: "hidden",
boxShadow: "inset 0 0 0 1px rgba(255,255,255,0.1), 0 8px 40px rgba(0,0,0,0.6)",
}}>
{/* Front camera dot */}
<div style={{
position: "absolute", top: "1.2%", left: "50%",
transform: "translateX(-50%)", width: "0.9%", height: "0.65%",
borderRadius: "50%", background: "#111113",
border: "1px solid rgba(255,255,255,0.08)", zIndex: 20,
}} />
{/* Bezel edge highlight */}
<div style={{
position: "absolute", inset: 0, borderRadius: "5% / 3.6%",
border: "1px solid rgba(255,255,255,0.06)",
pointerEvents: "none", zIndex: 15,
}} />
{/* Screen area */}
<div style={{
position: "absolute", left: "4%", top: "2.8%",
width: "92%", height: "94.4%",
borderRadius: "2.2% / 1.6%", overflow: "hidden", background: "#000",
}}>
<img src={src} alt={alt}
style={{ display: "block", width: "100%", height: "100%",
objectFit: "cover", objectPosition: "top" }}
draggable={false} />
</div>
</div>
</div>
);
}
iPad layout adjustments vs iPhone:
width: "65-70%" for iPad mockups (vs 82-86% for iPhone) — iPad is wider relative to its heightcanvasW (which is 2064 for iPad vs 1320 for iPhone)All sizing relative to canvas width W:
| Element | Size | Weight | Line Height |
|---|---|---|---|
| Category label | W * 0.028 | 600 (semibold) | default |
| Headline | W * 0.09 to W * 0.1 | 700 (bold) | 1.0 |
| Hero headline | W * 0.1 | 700 (bold) | 0.92 |
Vary across slides — NEVER use the same layout twice in a row:
Centered phone (hero, single-feature):
bottom: 0, width: "82-86%", translateX(-50%) translateY(12-14%)
Two phones layered (comparison):
Back: left: "-8%", width: "65%", rotate(-4deg), opacity: 0.55
Front: right: "-4%", width: "82%", translateY(10%)
Phone + floating elements (only if user provided component PNGs):
Cards should NOT block the phone's main content.
Position at edges, slight rotation (2-5deg), drop shadows.
If distracting, push partially off-screen or make smaller.
Dark/contrast background with app icon, headline ("And so much more."), and feature pills. Can include a "Coming Soon" section with dimmer pills.
html2canvas breaks on CSS filters, gradients, drop-shadow, backdrop-filter, and complex clipping. html-to-image uses native browser SVG serialization — handles all CSS faithfully.
import { toPng } from "html-to-image";
// Before capture: move element on-screen
el.style.left = "0px";
el.style.opacity = "1";
el.style.zIndex = "-1";
const opts = { width: W, height: H, pixelRatio: 1, cacheBust: true };
// CRITICAL: Double-call trick — first warms up fonts/images, second produces clean output
await toPng(el, opts);
const dataUrl = await toPng(el, opts);
// After capture: move back off-screen
el.style.left = "-9999px";
el.style.opacity = "";
el.style.zIndex = "";
toPng() loads fonts/images lazily. Second produces clean output. Without this, exports are blank.left: 0 before calling toPng.position: absolute; left: -9999px (not fixed).fontFamily on the offscreen container.01-hero-1320x2868.png, 02-freshness-1320x2868.png, etc. Use String(index + 1).padStart(2, "0").| Mistake | Fix |
|---|---|
| All slides look the same | Vary phone position (center, left, right, two-phone, no-phone) |
| Decorative elements invisible | Increase size and opacity — better too visible than invisible |
| Copy is too complex | "One second at arm's length" test |
| Floating elements block the phone | Move off-screen edges or above the phone |
| Plain white/black background | Use gradients — even subtle ones add depth |
| Too cluttered | Remove floating elements, simplify to phone + caption |
| Too simple/empty | Add larger decorative elements, floating items at edges |
| Headlines use "and" | Split into two slides or pick one idea |
| No visual contrast across slides | Mix light and dark backgrounds |
| Export is blank | Use double-call trick; move element on-screen before capture |
| User has no app screenshots yet | Tell them to take simulator screenshots first. Provide sizes: iPhone 15 Pro Max (1320x2868), iPad Pro 13" (2064x2752). Don't proceed without real screenshots. |
Present the final deliverable to the user:
## App Store Screenshots Complete
### Files Created
| File | Size | Device | Purpose |
|------|------|--------|---------|
| 01-hero-1320x2868.png | [size] | iPhone 6.9" | Hero slide |
| 01-hero-1284x2778.png | [size] | iPhone 6.5" | Hero slide |
| ... | ... | ... | ... |
### How to Upload
1. Open App Store Connect → Your App → App Store tab
2. Scroll to "Screenshots" section
3. Upload each size set to its corresponding device slot
4. iPhone 6.9" screenshots auto-apply to 6.7" and 6.3"
5. iPhone 6.1" is only needed for older device support
### Regenerating
Run the Next.js app (`bun dev`) and use the export buttons to regenerate at any size.
The generator is a permanent asset — update screenshots anytime by swapping PNGs in `public/screenshots/`.
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Add a new expert to the Remotion experts page
日本語の概要は準備中です。原文の説明を表示しています。
Add a new sound effect to @remotion/sfx
日本語の概要は準備中です。原文の説明を表示しています。
Use when setting up a secure email inbox for any AI agent — configuring inbound email via Resend, webhooks, tunneling for local development, and implementing security measures to prevent prompt injection attacks. Also use when someone mentions 'agent email', 'bot inbox', 'receive emails for agent', 'agent webhook', 'email security for AI', 'prompt injection via email', 'inbound email for bot', or wants their AI agent to receive and respond to emails securely.
日本語の概要は準備中です。原文の説明を表示しています。
Builds, measures, and improves an Answer Engine Optimization program across ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, and Copilot. Use for AEO, GEO, LLMO, AI search visibility, AI citations, brand mentions, answer-engine monitoring, llms.txt, AI crawler audits, or questions like 'what does AI say about us?' and 'why are competitors cited instead of us?'. Produces a durable evidence report and prioritized gameplan, not a one-shot content checklist.
日本語の概要は準備中です。原文の説明を表示しています。
Build detailed buyer personas and audience profiles from research. Use this skill whenever the user mentions audience, buyer persona, ideal customer, target market, ICP, watering holes, 'who am I selling to', customer research, or audience profile. Also use when content feels unfocused or generic (that's an audience problem), when conversion is low because messaging doesn't resonate, when starting any new project (audience should be first), or when any downstream skill needs audience.md but it doesn't exist yet. Even if the user doesn't explicitly ask for audience research, trigger this if they're writing copy or building a landing page without a clear audience defined. Three approaches: Quick Profile, Persona Build, Community Mining.
日本語の概要は準備中です。原文の説明を表示しています。
Agent eXperience Interface (AXI) router — the main entry point for agent-ergonomic CLIs. Use whenever the agent needs GitHub, browser automation, human review, quota routing, Slack, Notion, AWS/Docker/K8s, databases, GitLab (glab-axi), Oracle DB (oracle-axi), tmux/herdr/WezTerm sessions (cyber-mux), or any AXI-catalog tool; when choosing AXI vs MCP vs raw CLI; when building or reviewing an agent-facing CLI; or when the user says axi, gh-axi, chrome-devtools-axi, TOON output, agent ergonomics, or "prefer AXI". When in doubt about which external tool interface to use, start here.
日本語の概要は準備中です。原文の説明を表示しています。