Vibesku

VibeSKU CLI

Command-line interface for VibeSKU — AI-powered creative automation that turns product SKU photos into agency-grade e-commerce visuals and marketplace-ready copy in seconds.

What VibeSKU Does

VibeSKU replaces manual creative production for e-commerce teams. Upload a product photo, define brand guidelines, and receive complete multi-format visual sets. Core capabilities: VisionKV™ visual system (complete visual collections from a single product), high-fidelity product restoration (1:1 precision on textures/logos/packaging), multilingual text overlays, and omni-channel format output.

Credit costs: 1K/2K image = 1 credit, 4K image = 2 credits, short video = 5 credits, copy = 1 credit.

Setup

The CLI is bundled at bin/vibesku.js — self-contained Node.js script, zero dependencies (requires Node >= 18).

node <skill-dir>/bin/vibesku.js --help
alias vibesku="node <skill-dir>/bin/vibesku.js"   # optional

Skill Version

• Source of truth: VERSION
• Local version: cat VERSION
• Upstream repo: UllrAI/vibesku-agent

For version checks and auto-update workflow, see versioning.md.

Authentication

Two methods, resolved in priority order:

1. CLI token (vibesku auth login) — browser-based device flow, tokens at ~/.vibesku/config.json
2. API key — VIBESKU_API_KEY env var > config file key (from vibesku init or vibesku config set-key)

vibesku auth login                # Browser login (recommended)
vibesku init vsk_<key>            # API key for CI/CD

Template Selection Guide

VibeSKU provides 7 templates. Read the corresponding reference file before building the generate command.

Decision Tree

User wants visuals?
├── Technical exploded infographic → exploded-view
│   ├── Balanced callouts (default) → labelPlacement: balanced-callout
│   ├── Cleaner visual without labels → labelPlacement: none
│   └── Category-aware environment → backgroundMode: product-matched-scene
├── Existing image/poster needs visible text translated → image-translation
│   ├── Keep layout exactly → translationMode: faithful
│   ├── Make copy more natural → translationMode: polished
│   └── Preserve logos/trademarks → brandTextMode: preserve
├── Product photo needs a realistic usage scene → lifestyle-scene
│   ├── Let model infer scene → scenePreset: auto
│   ├── Specific setting → scenePreset: kitchen/outdoor/cafe/office/etc.
│   └── Human interaction → personMode: hand-only or full-figure
├── Marketplace/catalog white-background packshot → white-background
│   ├── Pure white background → backgroundTone: pure-white
│   ├── Softer depth → backgroundTone: soft-white
│   └── Tighter framing → cropMode: close-up
├── Single image (hero/banner/poster) → ecom-hero
│   ├── Product main photo → scenario: MAIN_IMAGE, aspectRatio: 1:1
│   ├── Marketing banner  → scenario: BANNER, aspectRatio: 16:9
│   └── Vertical poster   → scenario: POSTER, aspectRatio: 3:4
└── Multiple coordinated images → kv-image-set
    ├── Full detail page     → scenes: [kv-hero, lifestyle, detail-01, specs-table, ...]
    ├── Quick hero + lifestyle → scenes: [kv-hero, lifestyle]
    └── Brand storytelling   → scenes: [kv-hero, brand-story, user-review]

User wants text?
└── Product listing copy → listing
    ├── Amazon  → templateName: AMAZON_LISTING
    ├── Taobao  → templateName: TAOBAO_DETAIL
    ├── Shopify → templateName: SHOPIFY_DESC
    └── General → templateName: GENERIC

Quick Style Matching (image templates)

Typical Workflow

vibesku auth login                              # 1. Authenticate
vibesku templates                               # 2. Browse templates
vibesku templates info exploded-view            # 3. Inspect template details (optional)
vibesku credits                                 # 4. Check balance
vibesku generate -t ecom-hero \
  -n "Wireless Headphones" \
  -i product.jpg -l logo.png \
  -b "AudioTech"                                # 5. Generate visuals
vibesku status <job-id> --watch                 # 6. Monitor until complete
vibesku refine <output-uuid> -p "brighter bg"   # 7. Refine (optional)
vibesku export <job-id> -o ./output             # 8. Download results

Commands Quick Reference

All commands support --json for machine-readable output. Full details: commands.md

Key Behaviors

• Asset upload: -i auto-uploads product images; -l uploads logo
• Modify vs Regenerate: Use vibesku refine <output-id> -p "<instruction>" to edit existing output. Use vibesku generate only for new creations.
• Full UUID required for refine: Use vibesku status <job-id> --json to get complete output UUIDs (table view truncates them).
• Agent-friendly output: Prefer --json flag for all commands when used by AI agents.
• Version drift check (recommended triggers): Compare local VERSION with upstream when any of these happen: unknown template/option error, template mismatch with vibesku templates --json, user asks for newly added capability, or local check is older than 7 days (recommended cadence). If issues occur during execution, prioritize this check before retrying generation/refine commands. Follow versioning.md.
• Auto-update when outdated (recommended): If upstream version is newer, update skill first, then continue the user task with the updated skill.
• Watch mode: --watch polls every 5s until all runs complete.
• Credit tracking: Generation and refine show remaining balance on success.

Environment Variables