PackageFlow

Stop juggling terminal tabs. Start clicking.
_{A visual DevOps hub for npm scripts, Git, workflows, and deploy — controllable by AI via MCP.}

📥 Download • ✨ Features • 🤖 AI & MCP • 📚 Docs • 📸 Screenshots • ❓ FAQ

Why PackageFlow?

"Claude, deploy my feature branch." "Run tests and fix any failures." "Switch to the main branch and start the dev server."

One app to run scripts, manage Git, and deploy — across all your Node.js projects. The killer feature? Your AI assistant can control it via MCP.

Built with Tauri + Rust — fast, lightweight, 100% local.

🎬 Quick Start

# Install via Homebrew
brew tap runkids/tap && brew install --cask packageflow

Or download directly → Drag a project folder → Start clicking.

💡 Press Cmd + K for instant worktree switching

Try It in 60 Seconds

1. Import a project (drag a folder with package.json)
2. Click a script card like dev / build / test
3. Create a workflow for your repetitive tasks (dev → test → deploy)
4. (Optional) Enable MCP so your AI tool can run actions for you

Common Use Cases

• One-click “dev / build / test” across many repos
• Worktree switching without losing context
• Repeatable release pipelines (test → build → deploy)
• Fast security audits with AI explanations
• “AI as teammate” via MCP tool calls (safe + permissioned)

✨ Features

🤖 AI & MCP Integration

AI generates commit messages from your diffs

Multi-Provider AI Support

MCP Server — Let AI Control Your Dev Environment

PackageFlow exposes a Model Context Protocol (MCP) server that AI assistants can use:

"Claude, deploy my feature branch to staging."
"Run tests for all projects and report failures."
"Switch to the payment-fix worktree and start the dev server."

Works with:

• Claude Code
• Codex CLI
• Gemini CLI
• Any MCP-compatible AI tool

What AI Actually Does (MCP Tool Chains)

PackageFlow is “AI-driven” because the AI can call real tools (not just generate text). Example flows:

1) Understand a project

• You: “List my projects and show scripts for the frontend repo”
• Tools: list_projects → get_project

2) Run a script with context

• You: “Run tests for the packages/web project and summarize failures”
• Tools: get_project → run_npm_script

3) Generate a commit message from staged changes

• You: “Write a conventional commit message for what I staged”
• Tools: get_git_diff → (AI drafts message)

MCP Setup (Copy/Paste)

PackageFlow ships a companion MCP server binary: packageflow-mcp (stdio transport).

In PackageFlow, open Settings → MCP → MCP Integration and copy the generated config for:

• Claude Code / VS Code (JSON)
• Codex CLI (TOML)

Then your AI tool can call actions like list_projects, run_npm_script, run_workflow, and more.

AI CLI Integration

Run AI commands directly from PackageFlow:

You: "Analyze the security audit results and suggest fixes"
AI: Analyzing 3 high-severity vulnerabilities...

Security First:

• All API keys encrypted with AES-256-GCM
• Permission levels: Read Only → Confirm → Full Access
• Fine-grained tool permissions
• Complete request logging

Security & Privacy (Local-First)

PackageFlow is designed to keep your projects and secrets on your machine.

• Data stays local; AI features are opt-in
• Keys/tokens encrypted at rest
• MCP is permissioned (read → confirm → full)

Read more: Security & Privacy

📚 Documentation

Documentation home: English • 繁體中文 • 简体中文

📸 Screenshots

🎯 Projects + Scripts

🌳 Worktrees

⚡ Visual Workflow Builder

📦 Monorepo Support

🔗 Dependency Graph

🔀 Git Integration

🛡️ Security Audit

🚀 Deploy Accounts

💻 Terminals

🔌 Webhooks

🧳 Worktree Sessions

⌨️ Keyboard Shortcuts

📦 Installation

Homebrew (Recommended)

brew tap runkids/tap
brew install --cask packageflow

Upgrade

brew update && brew upgrade --cask packageflow

Troubleshooting

If you see an error like It seems the App source '/Applications/PackageFlow.app' is not there:

# Force uninstall the old cask record
brew uninstall --cask --force packageflow

# Reinstall
brew install --cask runkids/tap/packageflow

Direct Download

Download the latest release → Open the .dmg → Drag to Applications.

🗺️ Roadmap

Recently Shipped

• Multi-Provider AI — OpenAI, Anthropic, Google, Ollama, LM Studio
• MCP Server — Let AI assistants control PackageFlow
• AI CLI Integration — Claude Code, Codex, Gemini CLI
• AI Security Analysis — Intelligent vulnerability explanations
• Notification Center — Background task monitoring
• System Theme — Auto light/dark mode

Coming Soon

• 🪟 Windows Support — Cross-platform expansion
• 🐧 Linux Support — Complete desktop coverage
• 📦 Plugin System — Community extensions
• 🔄 MCP Actions — Custom AI-triggered workflows
• 🌐 Remote Collaboration — Team workflow sharing

💡 Request a feature or vote on existing ones!

❓ FAQ

🛠 Development

Prerequisites

• Node.js 18+
• Rust 1.70+
• pnpm

Setup

# Clone the repository
git clone https://github.com/runkids/PackageFlow.git
cd PackageFlow

# Install dependencies
pnpm install

# Start Vite (web UI)
pnpm dev

# Start the desktop app
pnpm dev:tauri

Build

# Build web assets
pnpm build

# Build the desktop app (dmg)
pnpm build:tauri

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Ways to Contribute

• 🐛 Report bugs and request features via Issues
• 🔧 Submit pull requests for bug fixes or new features
• 📝 Improve documentation
• 🔄 Share your workflow templates

Development Guidelines

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

🙏 Acknowledgments

Built with amazing open-source tools:

Tauri • React Flow • Lucide • Claude Code

If PackageFlow saves you time, give us a star!

_{MIT License • Made by runkids}