Auto-Claude: Turn Claude into a 12-Agent AI Dev Team That Works While You Sleep

Ever wished you could drop a project on Claude, walk away, and let it grind through tasks on its own—while you relax, play a game, or just wait out the workday? That’s exactly what Auto-Claude is designed to do.

Auto-Claude is an open-source desktop application that transforms Claude into a fully autonomous, multi-agent development system. Instead of a single AI assistant responding to prompts, you get up to 12 Claude agents running in parallel, each handling its own task, its own terminal, and its own isolated Git workspace.

If you’ve seen tools like Claude Engineer before, Auto-Claude takes that idea and pushes it much further.

From Claude Engineer to Auto-Claude: A Big Leap Forward

Claude Engineer already proved that Claude could work semi-autonomously through a Python-based CLI. Auto-Claude builds on that foundation—but upgrades everything around it.

What makes Auto-Claude different:

• A real desktop app, not just a CLI
• A visual Kanban board to manage tasks
• Up to 12 parallel Claude agents
• Each agent runs in its own terminal
• Every task is isolated using Git worktrees

In short, it feels less like “prompting an AI” and more like managing a small development team—except they don’t complain, don’t sleep, and don’t merge broken branches into main.

Parallel Agents Without Git Chaos

One of Auto-Claude’s smartest design choices is how it handles source control.

Each agent works inside its own Git worktree, meaning:

• No branch conflicts
• No accidental overwrites
• No tangled histories
• Clean, isolated task execution

When an agent finishes its work, you simply review, approve, and merge. No Git gymnastics required. It’s fast, clean, and surprisingly stress-free.

Installation Requirements

Before installing Auto-Claude, there’s one non-negotiable requirement:

• Claude Pro or Claude Max subscription

Auto-Claude relies on Claude Code, Anthropic’s official coding interface.

Step 1: Install Claude Code

If you don’t already have it:

npm install -g @anthropic-ai/claude-code

Step 2: Clone Auto-Claude

git clone https://github.com/AndyMik90/Auto-Claude.git
cd Auto-Claude

Step 3: Install Dependencies

Auto-Claude uses:

• Electron for the frontend
• Python for the backend

Install everything in one command:

npm run install:all

That’s it.

If you prefer not to build from source, precompiled binaries are also available for:

• Windows
• macOS (Intel & Apple Silicon)
• Linux (AppImage, .deb, and Flatpak)

Launching Auto-Claude (GUI and CLI)

Graphical Interface (Recommended)

Start the desktop app with:

npm start

This opens the Auto-Claude GUI, where you’ll find:

• A Kanban board for task management
• Live terminals for each agent
• Real-time progress monitoring

You create tasks, assign them to available agents, and let them run. Each terminal shows exactly what the agent is doing—live.

Command-Line Mode

If you prefer automation or scripting, Auto-Claude also supports CLI execution:

python run.py --spec 001

The number corresponds to a spec file located in:

specs/

You describe the task in plain language, and Auto-Claude handles the rest—from setup to execution.

How Auto-Claude Works Under the Hood

Auto-Claude uses a clean, modular architecture:

• Electron frontend
• Python backend
• Communication via WebSockets
• One isolated process per agent

When you launch a task, Auto-Claude automatically:

1. Creates a dedicated Git worktree
2. Spins up a Claude Code agent
3. Streams progress in real time
4. Runs QA and validation steps
5. Prepares the result for review and merge

All of this happens concurrently, across multiple agents, without stepping on each other’s work.

Why Auto-Claude Is a Big Deal

Auto-Claude isn’t just another AI wrapper. It’s a glimpse of what AI-driven development orchestration looks like when done right:

• True parallelism
• Strong isolation
• Clear visibility
• Minimal manual intervention

It’s especially compelling for:

• Large refactors
• Multi-feature builds
• Boilerplate-heavy projects
• Experimentation and prototyping

Final Thoughts

I’ve only started exploring Auto-Claude’s full potential, but even early testing makes one thing clear: this tool pushes Claude’s autonomy further than most projects out there.

If you’re excited by the idea of AI agents working independently, coordinated through a clean interface, Auto-Claude is absolutely worth your time.

It’s free, open source, runs on Windows, macOS, and Linux, and unlocks a whole new way of thinking about AI-assisted development.