You've been paying Cursor $20 to $60 a month. Or maybe you bailed for Zed last week, convinced the editor was the product and the AI was just a plugin bolted on top. The logic felt clean: pick the best-looking wrapper, swap models when a better one ships, stay neutral. Your IDE (integrated development environment — the app where developers write code) was the Switzerland of your workflow.

Except every "AI editor" rents the same four models from three labs. The wrapper is yours. The brain isn't. So the real question was always: what happens when the model vendor decides to ship its own wrapper?

On April 14, 2026, Anthropic answered. They shipped a complete Claude Code desktop redesign alongside Routines — and this wasn't a terminal tool with a fresh coat of paint. This is a full GUI (graphical user interface — the clickable windows you're used to, as opposed to a text-only command line) product. Multi-session sidebar, in-app file editor, rebuilt diff viewer for large changesets, integrated terminal, expanded preview pane for HTML and local servers, drag-and-drop layout. Three view modes. Side chat on ⌘+; so you can ask a branching question without polluting the main thread's context window (the AI's working memory — how much text it can "see" at once).

The kicker: according to The Register, the redesign is "competitive positioning — the company wants to prevent users from accessing Claude through VS Code plugins or third-party harnesses." Translation: Anthropic just stopped being a supplier 😼

Then there's Routines. Per Anthropic's own definition, a Routine is "a saved Claude Code configuration — a prompt, one or more repositories, and a set of connectors, packaged once and run automatically." Trigger it by schedule, API, or GitHub webhook (a URL that pings when something happens in your repo). Runs on Anthropic's cloud. Your laptop can be off. Daily caps: 5 on Pro, 15 on Max, 25 on Team and Enterprise. Overages billed per token.

Stack the pieces. CLI. Desktop. Routines. Managed Agents at $0.08 per session-hour. Skills. MCP (Model Context Protocol — a universal plug standard for AI tools, think USB but for feeding data to models). All first-party. All Anthropic. The model, the memory, the scheduler, the sandbox, the UI — one vendor, one invoice 🐈‍⬛

Not perfect. Claude-only — no GPT, no Gemini routing. macOS and Windows at launch; Linux is "in the following weeks" per VentureBeat's hands-on. No VS Code-class plugin ecosystem yet. And The New Stack put the warning in the headline: parallel sessions let you burn through tokens even faster. The lock-in didn't disappear. It just moved one landlord over 😹

Hands-on reviewers landed on the same line: if your workflow is "AI does the work, I review it," Claude Code Desktop is now the best option. If your workflow is "I write code, AI helps," Cursor still wins. That split is the whole story.

If you're evaluating an AI editor this week, the question stopped being "which editor is smartest." It became "which vendor do you want to owe everything to." Because the editor and the model now ship from the same company, and every future Cursor or Zed upgrade is a feature Anthropic has chosen not to ship yet. The neutral AI editor era ended on April 14, 2026 😼