PromptHub

A powerful tool for managing and organizing your prompts

PromptHub helps you better organize your prompts with comprehensive features including creating, editing, deleting, searching, and viewing prompts. Built with SwiftUI for macOS.

Apple Silicon (prebuilt binary):

brew tap leetaogoooo/prompthub https://github.com/LeetaoGoooo/PromptHub.git
brew install leetaogoooo/prompthub/ph

Intel, or to build from source with the Swift toolchain:

brew install --HEAD leetaogoooo/prompthub/ph

Verify the install:

ph --help

See the PromptHub CLI Package section below for usage and docs/cli-release.md for the full release & install reference.

• ✅ Create Prompt - Add new prompts with ease
• ✅ Edit Prompt - Modify existing prompts
• ✅ Delete Prompt - Remove unwanted prompts
• ✅ View Prompt - Browse and preview prompts
• ✅ Search Prompt - Find prompts quickly
• ✅ Status Bar - Quick access from menu bar
• ✅ Export/Import - Share prompts across devices
• ✅ AI Optimization - Enhance prompts with AI
• ✅ History - Track prompt usage
• ✅ Share Prompt - Collaborate with others
• ✅ Internationalization - Chinese & English support
• ✅ Prompt Gallery - Discover default prompts
• ✅ Prompt Diff - Compare prompt generated by ai
• 🆕 Single Prompt Testing - Test prompts across multiple AI models simultaneously
• 🆕 Model Comparison - Compare results from different AI models side-by-side
• 🆕 Real-time Streaming - Watch AI responses generate in real-time
• 🆕 Global Results View - View all model results in a comprehensive comparison interface
• 🆕 CLI Agent Skills - Install, configure, and manage skills for popular CLI agents (Claude Code, Cursor, Codex, Gemini CLI, etc.)
• 🆕 Granular Sandbox Access - Securely grant access only to specific CLI hidden folders (e.g. ~/.claude) without exposing your entire Home directory

PromptHub introduces a unified interface to manage context and skills for various AI CLI tools:

• Unified Skill Store: Discover and install open-source skills from skills.sh
• Multi-Agent Support: Compatible with Claude Code, Cursor, Codex, Gemini CLI, iFlow, Qwen Code, and Qoder
• Secure Sandbox Access: Granular control over which CLI configuration folders PromptHub can access, fully compliant with macOS App Store Sandbox restrictions
• Drafts & Previews: Edit skills securely in local drafts before publishing them to the CLI folders
• Project Scope: Install skills globally or restrict them to a specific project folder

The standalone CLI now lives in PromptHubCLI/ and exposes a real ph executable for exported prompts and agent skill operations.

App vs ph: use the app to author, edit, live-test, and grant agent folder access; use ph to script, automate, pipe JSON, and install skills from the shell or CI. Both surfaces read and write the same ~/.prompthub/ exports and agent skill directories.

• Build locally: swift build --package-path PromptHubCLI -c release --product ph
• Test locally: swift test --package-path PromptHubCLI
• Install with Homebrew (stable, Apple Silicon): brew tap leetaogoooo/prompthub https://github.com/LeetaoGoooo/PromptHub.git && brew install leetaogoooo/prompthub/ph
• Install with Homebrew (--HEAD, builds from source, also works on Intel): brew install --HEAD leetaogoooo/prompthub/ph
• Command examples: ph prompt list, ph prompt show, ph skill exports, ph skill install, ph skill list
• Release automation: pushing a ph-vX.Y.Z tag runs the GitHub Actions workflow that publishes ph-macos-arm64.tar.gz + .sha256 and smoke-installs Formula/ph.rb against the local archive before publishing

See PromptHubCLI/README.md for usage and docs/cli-release.md for the full release & install reference.

PromptHub now includes powerful prompt testing capabilities that allow you to:

• Select Multiple Models: Choose from all configured AI services and models
• Parallel Execution: Run tests across multiple models simultaneously for faster results
• Service Management: Easy configuration of different AI service providers

• Streaming Responses: Watch AI responses generate in real-time as they stream
• Live Updates: See results update dynamically across all selected models
• Progress Indicators: Visual feedback for ongoing tests

• Side-by-Side View: Compare outputs from different models in organized cards
• Global Comparison: Full-screen comparison view for detailed analysis
• Expandable Results: Collapse/expand individual results for better organization
• Full-Screen Mode: View individual results in dedicated full-screen windows

• Model Selection Sidebar: Intuitive interface for selecting and managing test models
• Test Controls: Easy-to-use controls for running tests and managing results
• Error Handling: Clear error messages and status indicators
• Copy to Clipboard: Quick copying of results for further use