A canvas ADE for human–AI collaboration: AI handles the complexity of learning and operating tools, so you can focus on goals, judgment, and intuitive interaction.

Work alongside Agents on one canvas, with visible progress and direct control whenever you need it.

Coding Agents, terminal tasks, long-running services, and real dependency connections in one branch workspace.

CleanCode.Multi-Agent.Collaboration.Demo.mp4

▶ Full demo (2:33, with sound)

CleanCode-.demo-10MB.mp4

🧩 What Is CleanCode?

CleanCode is a canvas ADE (Agentic Development Environment) for human–AI collaboration. It brings Agents, terminals, and project workspaces onto one canvas, so you can express intent, interact directly, and work with Agents on development and automation tasks.

Its design draws on System 1 and System 2 from Thinking, Fast and Slow: System 1 describes fast, intuitive responses, while System 2 involves focused analysis, planning, and reasoning. CleanCode applies these ideas to human–AI collaboration: people handle System 1 work, while Agents handle System 2 work.

Completing a task often means learning the software first: where to find an action, which parameters to set, and how the steps fit together. CleanCode aims to let Agents handle that learning and the steps that follow. Whether you are new to the workbench or already know it well, you can start with your goal, focus on the result, and take direct control whenever you want.

🖼️ Canvas Workbench: Arrange Agents, terminals, and groups on one canvas, with stacking, expansion, grid layout, and automatic organization.

🤖 Multiple Agents: Use 33 Agent CLIs, including Codex and Claude Code, with notifications from some tools when a reply finishes, input is needed, or approval is required.

🤝 Native Agent Collaboration: Use CleanCode MCP to let Agents operate the canvas, build workflows, create other Agents, and assign tasks.

🌿 Branch Workspaces: Work on multiple changes in parallel with Git worktrees, giving each branch its own canvas, terminals, and Agent sessions.

📝 Issue-Driven Development: Browse GitHub Issues and create or open a branch workspace linked to an issue.

🧰 Workspace Presets: Set default Agents and terminal templates for a project, add them to new branch workspaces automatically, and optionally run the templates.

⚙️ Automated Workflows: Connect terminal commands, scripts, and non-interactive AI CLIs, run them according to their dependencies, and view the state of every step.

🔌 Port Management: Choose fixed, preferred, or automatically assigned ports for services running in parallel, and see their actual addresses.

📦 Templates and Quick Execution: Save frequently used terminals, workflows, or groups as templates to avoid repeated setup, and launch common tasks with keyboard shortcuts.

See Product Features and Quick Start for details.

CleanCode is built on terminals and includes launch configurations for 33 Coding Agents, with automatic detection of tools installed on your machine so you can use them in the current workspace.

Other Agents that run from the command line can also run directly in a canvas terminal, using your familiar tools and configuration.

Keep using the Agents you know and collaborate on one canvas.

If you just want to get CleanCode running, you do not need to understand every concept first. The steps below are enough.

Download the installer for your platform from GitHub Releases:

The current Preview is not signed with official developer certificates. Download it only from this repository's GitHub Releases, and verify the package against SHA256SUMS.txt from the same release.

Because the app has not yet completed Developer ID signing and Apple notarization, macOS will block it the first time it opens. After moving CleanCode.app to /Applications and verifying the source and SHA-256 checksum, you can remove the download quarantine attribute for this app only and launch it:

xattr -dr com.apple.quarantine /Applications/CleanCode.app open /Applications/CleanCode.app

This command bypasses the initial Gatekeeper check for this app. Do not replace the target with a broad path such as /Applications, your Downloads folder, or your home directory. Alternatively, follow Apple's official instructions and choose Open Anyway under System Settings → Privacy & Security.

These requirements apply only when running from source or contributing to development. You do not need to install these development dependencies before downloading the Preview.

git clone https://github.com/chen-985211/cleancode.git cd cleancode pnpm install pnpm dev

Common Checks

pnpm typecheck pnpm test pnpm pre-commit

Build Locally

Unpacked app for the current platform, intended for local verification

pnpm package

Distribution installer for the current platform

pnpm dist

You can also select a target explicitly on the corresponding operating system

pnpm dist:mac pnpm dist:win pnpm dist:linux

Build artifacts are written to release/.

Product Features and Quick Start

Documentation Center

Architecture Guide

Context Map

UI Contract

Terminal Dependency Workflows

Agent and Session Lifecycle

Native CleanCode MCP

Development Guidelines

🤝 Contributing

Issues, discussions, and pull requests are welcome. Before you begin, read the Contributing Guide and Development Guidelines.

Scan the code below to join the CleanCode community, share ideas, suggest features, or report problems, and help make CleanCode better.

Group number: 186885114

Explore the possibilities of an AI-native operating system through human–AI collaboration.