DUMPSTER
leescott2012/dumpster
Curação inteligente de fotos e sequenciamento para carrosséis Instagram perfeitos
Ver no GitHubFerramenta de curação e sequenciamento de fotos alimentada por IA que transforma bibliotecas desorganizadas em carrosséis perfeitos para Instagram. Usa clustering inteligente, heurísticas de sequenciamento proprietárias e geração de legendas com múltiplos provedores de IA para criar narrativas visuais coesas.
Instalação
npx convez-ai leescott2012/dumpster1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Criadores de conteúdo organizando foto dumps para redes sociais
- Influenciadores e profissionais construindo carrosséis visualmente coesos
- Otimização automática de sequência de fotos com análise estética
Veredito

Plugin sólido para criadores que precisam profissionalizar foto dumps em Instagram, com stack técnico robusto (SwiftUI, multi-provider LLM). Útil especialmente para quem opera em iOS, embora a proposta seja nicho e fortemente orientada a redes sociais.
README
A tradução para o português deste README ainda está na fila.
DUMPSTER
DUMPSTER is a high-end photo curation and sequencing platform designed to help creators build the perfect Instagram "dump." By combining aesthetic heuristics with advanced AI, Dumpster transforms a chaotic photo library into a curated narrative.
🚀 Executive Summary
In the era of the "photo dump," the challenge isn't taking photos—it's selecting and sequencing them. Dumpster uses a proprietary Dump Formula to score and arrange photos based on visual contrast, category mix, and narrative flow. Whether you're a professional creator or a casual user, Dumpster provides the tools to build carousels that feel intentional and high-end.
✨ Core Features
🧠 AI-Powered Curation
- Intelligent Clustering: Automatically groups photos from your library into suggested "dumps" using Apple Vision and custom clustering algorithms.
- Automated Sequencing: Uses the "Dump Formula" to arrange photos into a 7 or 12-slot carousel, ensuring a perfect mix of "hooks," "details," and "atmosphere."
- Multi-Provider Captions: Generates professional, on-brand captions using Anthropic Claude, OpenAI, and other LLM providers.
📱 Dual-Platform Architecture
- Native iOS App: A high-performance SwiftUI implementation leveraging SwiftData for local persistence and Apple Vision for on-device image analysis.
- Web Application: A modern React + Vite + TypeScript web app for rapid curation and cross-platform accessibility.
🛠 Professional Workflow
- Vibe Check: Real-time analysis of color temperature and category consistency across your dump.
- Undo/Redo System: A robust, snapshot-based state management system for non-destructive editing.
- Custom Heuristics: Fine-tune the arrangement logic to match your personal aesthetic.
🛠 Tech Stack
iOS (Native)
- Framework: SwiftUI
- Persistence: SwiftData
- AI/ML: Apple Vision (Image Classification & Clustering)
- Networking: Multi-provider LLM integration (OpenAI, Claude, Gemini, etc.)
Web
- Framework: React 19 + Vite
- State Management: Zustand
- Styling: Modern CSS with native-feel design tokens
- Drag & Drop: @dnd-kit
🗺 Roadmap
The project is currently undergoing a Native Conversion, migrating the core web UI into a 100% native SwiftUI experience to leverage deeper iOS integrations and improved performance. See docs/NATIVE_CONVERSION_SPEC.pdf for the full technical specification.
Built for the next generation of digital storytellers.