A library for managing, sharing, and discovering AI prompts.
Stack
- Database: PostgreSQL 16 (Docker)
- Backend: Node.js + Express + TypeScript + Drizzle ORM
- Frontend: React + Vite + TypeScript
Core
- Store, organize, and manage AI prompts with full CRUD operations
- Folder-based organization with nested hierarchies
- Tag system with bulk rename, merge, and management
- Version history with one-click restore
- Favorites and rating system
- Grid and list view modes
Search
- Full-text keyword search across titles, content, and descriptions
- Semantic search powered by pgvector embeddings, find prompts by meaning, not just keywords
- Faceted filtering by status, type, model, tags, and folders
AI-Powered
- Prompt structure analysis with type detection and confidence scoring
- AI-generated improvement suggestions
- Streaming prompt improvements with accept/discard workflow
- Interactive playground, test prompts against Claude, GPT, Gemini, Grok, and more
Prompt Types
- System, Task, Template, Chain, Reference, and Snippet classifications
- Auto-detection of prompt type
- Template variable extraction with {{variable}} substitution and live preview
Multi-Model Support
- API key management for Anthropic, OpenAI, Google AI, xAI, and OpenRouter
- Model compatibility tagging per prompt
- Playground with configurable temperature and token limits
Infrastructure
- Docker Compose deployment (Express + PostgreSQL + pgvector)
- OIDC authentication support
- OpenAPI/Swagger documentation
- Dark/light theme with system preference detection
- Responsive design with View Transitions API
