app-media-creator

skill
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Uyari
  • fs module — File system access in .claude/hooks/ui-workflow-guard.cjs
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Generate every App Store, Google Play and web media asset from a logo, screenshots and a video, in six languages.

README.md

app-media-creator

Turn a logo, screenshots and a video into every media asset your app needs for the
Apple App Store, Google Play and the web — in six languages, from one local tool.

Status: early development (pre-0.1.0). The project setup is done; features are built milestone by
milestone (see Roadmap). Nothing is published to npm yet.

What it will generate

Input Output
Logo App Store icon 1024 (+ dark / tinted, layer-ready files for Icon Composer), Google Play icon 512 and adaptive icon layers with mipmaps, favicon set and site.webmanifest, Open Graph / X / WhatsApp previews, social profile and banner sizes, monochrome and inverse variants, labelled icon variants for preview and test installs (docs), and Expo / React Native icons with their app.json lines (docs)
Screenshots Store screenshots with headline card, background and device frame (3 layouts) — or frame-only images for your website; all iPhone (incl. iPhone Duo), iPad and Android phone / tablet sizes; App Store product page header and search result assets; Google Play feature graphic
Video, URL or repo App Store app previews (886×1920, 15–30 s, spec-checked), looping header / search videos, promo reels for Google Play, web and social, or your video embedded in a styled background
AI copy (optional) Screenshot headlines that tell a story, and store texts (name, subtitle, description, keywords …) ready to paste

Every text-bearing asset is generated per language — one, several or all of:
German de-DE, English (US) en-US, Spanish es-ES, Italian it-IT, French fr-FR, Turkish tr-TR.
Everything downloads as one ZIP in a store-ready folder layout.

How AI is used

AI is optional and never needs your credentials inside the tool:

  • Default: the UI prepares a job folder; you run the bundled Claude Code skill in your editor, which uses
    your own Claude Code login. The UI picks up the results automatically.
  • Optional: put an ANTHROPIC_API_KEY in .env to get a "Generate" button directly in the UI.

Screenshot headlines with Claude Code

  1. Install the skill once in the project that uses app-media-creator:

    npx skills add halukmt/app-media-creator --skill amc-copy
    
  2. In the Screenshots section, open Preview and edit and choose Prepare AI job for the selected languages.

  3. Copy the command exactly as the page shows it (for example
    /amc-copy "/path/to/your-app/.app-media/jobs/20261119-104512-k3x9", with the full path) and run it in Claude Code. The skill looks at your screenshots, follows the copy rules and your claims (allowed and forbidden),
    and writes two variants per language: A outcome-first and B mechanism-first.

  4. The page shows the result as soon as the skill has written it. Compare A and B and apply one; nothing changes
    before you do. Results that break the rules (for example a forbidden claim) are shown as errors and never applied.

Store texts with Claude Code

  1. Install the skill once: │
    ● claude-code_2-1-327_agent Agent detected — installing non-interactively
    [?25l│
    ◇ Source: https://github.com/halukmt/app-media-creator.git
    [?25h[?25l│
    ◒ Cloning repository…◐ Cloning repository…◓ Cloning repository…◑ Cloning repository…◒ Cloning repository…◐ Cloning repository…◓ Cloning repository…◑ Cloning repository…◒ Cloning repository….◐ Cloning repository….◓ Cloning repository….◑ Cloning repository….◒ Cloning repository….◐ Cloning repository….◓ Cloning repository….◑ Cloning repository….◒ Cloning repository…..◐ Cloning repository…..◓ Cloning repository…..◑ Cloning repository…..◒ Cloning repository…..◐ Cloning repository…..◓ Cloning repository…..◑ Cloning repository…..◒ Cloning repository…...◐ Cloning repository…...◓ Cloning repository…...◑ Cloning repository…...◒ Cloning repository…...◐ Cloning repository…...◓ Cloning repository…...◑ Cloning repository…...◒ Cloning repository…...◐ Cloning repository…◓ Cloning repository…◑ Cloning repository…◒ Cloning repository…◐ Cloning repository…◓ Cloning repository…◑ Cloning repository…◒ Cloning repository…◐ Cloning repository….◓ Cloning repository….◑ Cloning repository….◒ Cloning repository….◐ Cloning repository….◓ Cloning repository….◑ Cloning repository….◒ Cloning repository….◐ Cloning repository…..◓ Cloning repository…..◇ Repository cloned
    [?25h[?25l│
    ◒ Discovering skills…◐ Discovering skills…◓ Discovering skills…◑ Discovering skills…◒ Discovering skills…◇ Found 13 skills
    [?25h│
    ■ [SKILLS_NO_MATCHING_SKILLS] No matching skills found for: amc-store-texts│ ╰▶ fix: Available skills:
    │ - amc-copy
    │ - next-issue
    │ - openspec-apply-change
    │ - openspec-archive-change
    │ - openspec-explore
    │ - openspec-propose
    │ - openspec-sync-specs
    │ - openspec-update-change
    │ - release
    │ - daisyui-colors
    │ - daisyui-config
    │ - daisyui-install
    │ - daisyui-usage.
  2. In the Store texts section, choose the languages and Prepare AI job; run the command the page shows in Claude Code.
  3. The skill learns what the app does (screenshots, current texts, your repo's README), writes name, subtitle,
    promotional text, description, keywords and what's new for the App Store and title, short and full description for
    Google Play, within each store's limits, and checks them. Keywords are researched per language, not translated.
  4. The page shows the texts as soon as they are written; apply them, then edit any field. Counters show each field's
    limit (Apple keywords in bytes), and the export packs the texts per store and language in paste order.

Requirements

  • Node.js ≥ 22.12
  • ffmpeg and ffprobe on PATH (or FFMPEG_PATH / FFPROBE_PATH in .env) for video features. Without them
    everything else works; video generation says what is missing, and the export checklist cannot measure videos.

App Store preview videos

POST /api/generate/store-previews turns each uploaded screen recording (up to three per language) into App Store
previews: 886×1920 for iPhone (also used by iPhone Duo, experimental) and 1200×1600 for iPad when the project picked
an iPad size; a landscape recording gets the landscape size. The recording is trimmed (trim), padded to the store size
(never cropped), faded in and out and framed by optional text cards per language (intro, outro, cardSeconds).
Output: H.264 at 30 fps and 11 Mbps, stereo AAC (silence when the recording has no audio), 15–30 s, checked with ffprobe.
App Store Connect sets the poster frame per preview (default: the frame at 5 s); GET /api/video/frame shows a frame
of a generated preview to choose it. Only real app footage belongs in a preview.

Device videos and looping header / search videos

POST /api/generate/embed-videos puts each recording in a device frame on a background for websites and social
posts: 9:16 (1080×1920), 16:9 (1920×1080) and 1:1 (1080×1080), with an optional line of text per language (texts),
frame, background (default: the header image background) and trim. The recording's audio is kept.

POST /api/generate/loop-videos makes the App Store's looping product page header (3840×1646) and search results
(1920×1280) videos: the layout of the header and search images with the language's phrase, and the recording playing
in the key-art device. The end crossfades into the start, so the loop has no jump; 30 fps, 5–30 s, no audio (the App
Store plays them muted).

Local development

npm install
cp .env.example .env   # optional
npm run dev            # API server on :4790 + UI on http://localhost:5173
Command Purpose
npm run dev Server (watch mode) and Vite UI together
npm run build Build UI (dist/web) and server / CLI (dist/)
npm start Run the built CLI (app-media-creator)
npm run check TypeScript checks for server, UI and tests
npm run check:all Project checks (component inventory, …)
npm test Unit tests (Vitest); see docs/testing.md for a faster local run
npm run test:e2e End-to-end tests with accessibility checks (Playwright + axe)
npm run check:release Release gate: CHANGELOG version equals package.json, no [Unreleased] entries

CLI options: app-media-creator --port <n> --no-open --version --help.

Roadmap

Milestone Content
M0 Project setup (this)
M1 Core, app shell, locale settings, logo pack
M2 Screenshots: card mode, frame-only mode, all devices, per-locale sets
M3 AI copy: headline storytelling, job folder, Claude Code skill
M4 App Store header / search result assets, Play feature graphic, in-app event media
M5 Video: store previews, looping header video, promo reels (HyperFrames or hype-reel)
M6 Store texts for all locales
M7 First release 0.1.0, npm package, public repo

How we work

  • Every requirement is a GitHub issue in the project board, with size:*, prio:* and enhancement / bug labels.
  • Implementation is planned with OpenSpec in
    openspec/changes/<issue>-<slug>/.
  • Feature branches merge into dev via PR; releases go from dev to main.
  • Changes are recorded in CHANGELOG.md (Keep a Changelog, SemVer).
  • Rules for contributors and AI agents: .github/copilot-instructions.md.

Licence

MIT. Store specifications are summarized from Apple and Google documentation; Apple device bezels
and badges are not included and must be downloaded from Apple by each user.

Yorumlar (0)

Sonuc bulunamadi