Install

System requirements

Requirement	macOS	Windows	Linux
OS version	macOS 13+ (Ventura)	Windows 10+	A modern distro with a graphical environment
Architecture	Apple Silicon (arm64) and Intel (x86_64)	x86_64	x86_64
Disk space	~200 MB for the app + space for worktrees	~200 MB	~200 MB
Git	Required	Required	Required

Download

macOS is the shipping platform today. Windows and Linux builds are in progress; you can build from source on all three platforms in the meantime.

macOS: Download the latest .dmg from GitHub Releases, open it, and drag Control Center to Applications. The build is signed with a Developer ID and notarized by Apple, so it opens normally. You can also install it with Homebrew:
Terminal window
```
brew tap control-center/tap
brew install --cask control-center
```
Windows and Linux: prebuilt downloads are not available yet. Watch GitHub Releases for the first Windows .exe and Linux AppImage, or build from source today.

Agent CLI setup

Control Center dispatches agents through CLI backends. Install at least one:

Claude Code (claude): npm install -g @anthropic-ai/claude-code
Codex CLI (codex): npm install -g @openai/codex
Pi (pi): Install per the Pi documentation

Control Center auto-detects installed CLIs in your PATH. You can also register custom adapters in Settings → Adapters.

GitHub authentication

Control Center needs a GitHub personal access token (PAT) with these scopes:

repo: full repository access
read:org: organization membership

Create one at GitHub Settings → Developer settings → Personal access tokens.

If you have the gh CLI installed and authenticated (gh auth login), Control Center detects and uses it automatically.

Choose how Control Center runs

Control Center is a thin client: it never opens the database itself — a headless cc_server process owns the data and serves it over RPC. On first launch you choose how the desktop reaches its server:

Run locally (default): the desktop spawns and supervises a cc_server on this machine that owns the database, and talks to it over loopback. This is the self-contained, single-machine setup — nothing to configure beyond the choice.
Connect to a remote server: point the desktop at a cc_server running elsewhere (a small box, a cloud VM) with the server’s WebSocket URL, a device id, and the pairing key it minted for you. The data lives on that server; this desktop is purely a renderer.

You can change this later in Settings → Server connection (it applies on the next restart). The web client is always remote — a browser cannot run a server.

Run a headless server

To run your own cc_server for other clients to connect to, build the pure-Dart binary and pair a client before first start. See Run a headless server for the full walkthrough; the essentials:

cd apps/cc_server
dart build cli
./build/cli/<os_arch>/bundle/bin/cc_server pair --data-dir ./data --port 9030
./build/cli/<os_arch>/bundle/bin/cc_server --data-dir ./data --port 9030

Web client

The web build is the same app compiled for the browser. It always connects to a remote cc_server — open the deployed client, enter the server’s WebSocket URL, device id, and pairing key, and it renders the full desktop interface. You can also build it yourself with flutter build web.

Verify your setup

After launching Control Center:

The first-run flow asks how you want to run (local or remote), then walks you through GitHub auth and workspace creation.
Open Settings → Adapters to verify your agent CLIs were detected.
Open Settings → Repositories to add your first repo.
Run Agent doctor from any agent’s context menu to diagnose environment issues.

Building from source

git clone https://github.com/SamuelAlev/control-center.git
cd control-center
flutter pub get
flutter pub run build_runner build --delete-conflicting-outputs
flutter run -d macos  # or -d windows, -d linux

Requires Flutter 3.x with desktop support enabled.