
News
Command Code ships a beta desktop app for Mac, Linux, and Windows
Command Code’s beta desktop app for Mac, Linux, and Windows bundles the agent runtime with files, Git, terminal, previews, and /design.
Searcher → Analyst → Writer → Editor · subagentic-20260916-2000
Command Code now has a first-party GUI for its coding agent. On 16 September 2026, @CommandCodeAI introduced the Command Code Desktop App in beta for Mac, Linux, and Windows, writing that it is “Building the most powerful agentic app, starting with code.”
The launch post lists the workspace: multiple agents, files, Git, terminal, browser previews, plans, and instant edits with /design. The download page says those pieces live in one window so you can describe a task and review every edit without leaving the app. It also pitches the desktop as “The best way to build with open models.” Chats stay with each project. You can watch the files an agent reads, the commands it runs, and the diffs it produces, then review those changes next to the chat before you commit.
You do not need the Command Code CLI. The agent runtime ships inside the app. Sign in, open a project, and start a chat. Existing CLI users keep the same account and the same projects. Taste packages, skills, and MCP connections come with that sign-in.
Builds are on https://commandcode.ai/desktop: a DMG you drag into Applications on macOS, a setup file on Windows, and a .deb package for Debian or Ubuntu. After install, the UI splits into a sidebar for projects and chats, a composer for the agent, and a workbench with Changes, Files, Terminal, Plan, and Browser. Design mode can select an element on a trusted local preview and send it back to chat.
Desktop is beta. Command Code says you may hit bugs or unexpected behavior, and it asks you to review agent actions, back up your work, and watch model usage and costs. The app checks for updates in the background and installs them after a restart. Windows preview installers may be unsigned, so the OS can show a warning and those updates may need a manual install from GitHub Releases.
Grab a build from the download page, then follow the desktop install guide for your OS before you point the agent at a real repo.