Claude Code & Agent Skill to generate Upload-ready App Store screenshots that drive installs — powered by GPT-Image 2.
video.mp4
Type /shots in Claude Code. Paste your App Store URL to scrape your listing,
let the skill research your app, market, theme, colors, and competitors,
add your own screenshots to .shots/app-screenshots/, drop inspo into .shots/inspo/,
and get scroll-stopping App Store screenshots in minutes — not days.
⚠️ API Key Required You need either:
OPENAI_API_KEYwith image generation enabled (GPT Image 2 access), ORFAL_KEYfor fal.ai generationOptional:
SEARCHAPI_KEYfrom SearchAPI.io (free trial available) enables scraping your or other apps' existing App Store screenshots to use as reference.
Install the Shots skill to your AI coding agent:
# Install to your current project
npx skills add hypersocialinc/shots
# Install globally (available in all projects)
npx skills add hypersocialinc/shots --global
# Install to specific agents
npx skills add hypersocialinc/shots --agent claude-code cursorWorks with Claude Code, Cursor, Codex, Cline, and 40+ other AI coding agents.
![]() |
![]() |
![]() |
| TikTok — bold claims, scroll-stopping color | ||
![]() |
![]() |
![]() |
| Miniparty — playful identity, clear value props | ||
![]() |
![]() |
![]() |
| Snapchat — high-energy visuals, benefit-first copy | ||
![]() |
![]() |
![]() |
| ChatGPT — clean product tour, clear benefit headlines | ||
| Platform | Status | Composite Size | Final Panel Size |
|---|---|---|---|
| iPhone | ✅ Fully tested | 3456×2400 | 1284×2778 |
| iPad | 6144×2732 | 2048×2732 | |
| Android Phone | 3240×1920 | 1080×1920 | |
| Translation | - | - | |
| Apple Watch | ❌ Not yet supported | - | - |
| App Store Videos | 🔜 Planned | - | - |
~$0.50 per generation (3 screenshots) · ~$0.17 per screenshot
Output image tokens account for ~95% of the cost and are fixed per composite size — prompt length and reference images are negligible. Based on real generation data and OpenAI GPT Image 2 pricing:
| Tokens | Cost | |
|---|---|---|
| Input (text + images) | ~1K-6K | $0.005-0.04 |
| Output (composite image) | 16,554 | $0.497 |
| Total | ~$0.50 |
Everything runs through a single /shots command. Sub-flows are triggered conversationally:
| Input | What it does |
|---|---|
/shots |
Full creation flow — questionnaire, workspace setup, generate screenshots |
/shots <url> |
Scrape the App Store URL, then start creation flow |
/shots revise |
Iterate on existing shots with targeted feedback |
/shots translate [shot-id] [locale] |
Localize an existing shot/run into another language/locale |
/shots scrape <url> |
Scrape App Store metadata only (no generation) |
You can also trigger sub-flows mid-conversation by saying "revise", "translate to Japanese", "scrape", etc. Benefit headlines are crafted from a saved strategy brief during the create flow.
Shots is a research-first screenshot generation system that produces conversion-optimized App Store and Google Play assets:
Sets up a .shots/ directory in your project with config schema, manifest tracking, and folders for reference images.
Automatically gathers context from multiple sources:
- App Store scraping — Pulls metadata, ratings, description, and existing screenshots from your listing (iTunes API + optional SearchAPI for device-grouped screenshots)
- Repository analysis — Reads your README, package.json, and source code to understand your product
- Visual references — Inspects screenshots you drop into
.shots/app-screenshots/and inspiration images in.shots/inspo/ - Competitor mapping — Identifies 3-5 direct competitors and analyzes their positioning, visual themes, and review language
SearchAPI Integration: When SEARCHAPI_KEY is set, the scraper fetches device-grouped screenshots (iPhone, iPad, etc.) from the App Store, providing richer visual references than the iTunes API alone. Sign up at searchapi.io — free trial available.
Saves a reusable research artifact (strategyBrief in config.json) containing:
- App summary, target audience, market maturity
- Positioning (core promise, differentiator, proof points)
- Competitor analysis with visual/copy patterns
- Voice and language bank (tone, market words, taboo phrases)
- Visual theme (style family, mood, palette, motifs)
- Panel arc structure (Value → Flow → Trust narrative)
Generates 6-8 problem-first benefit statements ranked for install intent (transformations, not features). Headlines are saved to config.json and reused across revisions/translations.
Creates wide panoramic composites via OpenAI's GPT Image 2 (direct API or fal.ai) using:
- Saved strategy brief and benefits (not regenerated each time)
- Reference images for visual consistency
- Platform-specific dimensions (iPhone: 3456×2400, iPad: 6144×2732, Android: 3240×1920)
- Verbatim quote-based prompting for >95% text accuracy
Automatically splits composites into 3 upload-ready individual screenshots with crop-safe boundaries (no text cutoff or visible seams).
- Revise: Make surgical changes to existing shots while preserving strategy and visual identity
- Translate: Localize screenshot copy into other languages/locales without rebuilding visuals
All generated assets are versioned in timestamped run directories under .shots/runs/ (gitignored).
- Node.js 18+
OPENAI_API_KEYorFAL_KEYSEARCHAPI_KEY(optional) — improves App Store screenshot scraping with device grouping
skills/shots/
SKILL.md # Router, shared setup, config contract
reference/
create.md # Create flow
strategy.md # Research + strategy brief + benefits
prompting.md # GPT Image 2 prompt contract
revise.md # Revision flow
translate.md # Localization flow
scrape.md # App Store scrape flow
scripts/ # generate.mjs, crop.mjs, scrape.mjs
.shots/ # Per-project workspace (created at runtime)
config.json # Scaffold-generated app/workflow schema
runs/ # Generated output (gitignored)
Using Shots for your app? We'd love to see it! Open a PR adding a link to your App Store or Google Play listing and we'll feature it here.












