Claude Code gets a major desktop redesign for parallel agents

If you've been using coding agents heavily, you know the job doesn't feel like typing a prompt and waiting around anymore. You're kicking off a refactor over here, fixing an edge-case bug in another repo, and running a test suite in a third, acting more like an engineering lead reviewing work than a typist.

Anthropic clearly sees that shift. The company just released a complete redesign of the Claude Code desktop app, rebuilding the entire workspace around running parallel agentic workflows.

The multi-repo conductor seat

The centerpiece of the new app is a dedicated sidebar built to keep tabs on multiple active and recent sessions at once. You can spin up tasks across different repositories, jump between them as results stream in, and filter or group sessions by status, project, or environment. Once a pull request merges or closes, the session archives itself so your workspace stays tidy.

One particularly smart addition is side chat. By hitting a quick shortcut, you can branch off a temporary conversation to ask Claude a question mid-task. The side chat inherits context from the active task, but it never feeds back into the main thread, meaning you won't accidentally steer a running agent off course.

Everything in one screen

Anthropic also packed in several core development tools so you can stop bouncing between your full-blown editor and the app. There's now an integrated terminal for running local tests and builds, an in-app file editor for making fast spot tweaks, and a rebuilt diff viewer optimized for large changesets.

The preview pane has been expanded to support local servers alongside direct rendering of HTML files and PDFs. The entire layout is now drag-and-drop, meaning you can pin your diff viewer, terminal, and chat panes in whatever grid layout fits your screen.

Under the hood, the app gains feature parity with CLI plugins, and remote SSH support now covers Mac alongside Linux. You can toggle between three view modes depending on how much tool-call detail you want to see, and a new usage button keeps your context window and session limits visible. The update is live now for Pro, Max, Team, and Enterprise users, as well as via the Claude API.

Why it matters

This isn't just a UI refresh. It marks a fundamental shift in how developers interact with AI agents: moving from sequential turn-taking to parallel orchestration.

Most AI coding tools still assume you're sitting in an IDE waiting on one prompt to finish. Anthropic's bet is that real engineering work happens across multiple repositories at once, and the biggest friction point is context pollution and session management. Side chats and auto-archiving PRs are pragmatic features that could only come from watching engineers use these tools all day.

For power users already paying for Claude tiers, it's definitely worth checking out today, especially if you spend half your afternoon orchestrating multi-repo tasks. The main catch will be token spend: running three or four active coding agents in parallel burns through context and quota at an alarming pace, and you'll want to watch that new usage meter closely.