Planning a complex code change is hard enough. Reviewing it in a terminal window shouldn’t make it harder.
Anthropic is addressing that friction with a new capability called Ultraplan, currently in research preview as part of Claude Code. The feature moves the planning phase of a coding task from your local terminal to the cloud — and gives developers a richer environment to review, revise, and approve a plan before a single line of code changes.
It’s a small workflow shift with real practical value, especially for teams working on large-scale migrations, service refactoring, or anything that requires careful coordination before execution begins.
How it Works
Ultraplan connects Claude Code’s command-line interface (CLI) to a cloud-based session running in plan mode. When a developer triggers it — either by running /ultraplan followed by a prompt, typing the word “ultraplan” anywhere in a standard prompt, or choosing to refine an existing local plan in the cloud — Claude picks up the task remotely and starts drafting.
While the cloud session is running, the developer’s terminal remains free. A status indicator tracks what’s happening: Claude is researching the codebase, Claude needs clarification, or the plan is ready for review. That last status, ◆ ultraplan ready, is the signal to open a browser and take a look.
The review experience is notably more useful than scanning output in a terminal. Developers can highlight specific passages and leave inline comments, use emoji reactions to flag sections they approve or want revised, and navigate a structured outline sidebar to jump between parts of the plan. Claude addresses the comments, revises the plan, and presents an updated draft. The cycle continues until the plan looks right.
Two Paths to Execution
Once a developer approves the plan, Ultraplan offers two options for what happens next.
The first is cloud execution. Claude implements the plan in the same web session where it was reviewed. The terminal gets a confirmation, the status indicator clears, and the work continues remotely. When implementation is complete, developers can review a diff view and open a pull request — all from the browser.
The second option is what Anthropic calls “teleport back to terminal.” The approved plan returns to the local environment, where the developer has full access to their own tools, configurations, and dependencies. This is the better option when local execution matters — integration environments, internal tooling, anything that doesn’t live in the cloud.
Back in the terminal, the developer has three options: Inject the plan into the current conversation and continue from there, start a new session with the plan as the only context, or cancel and save the plan to a file for later review.
It’s a clean handoff in either direction.
Why This Matters for DevOps Teams
The friction between planning and doing is one of the more underappreciated pain points in software development. Teams spend real time trying to align on what a change will involve before they start making it — and the tools available for that alignment often fall short. A wall of terminal output is not a great review surface.
Ultraplan gives that review step a proper home. Inline comments mean feedback is targeted and specific, not a general response to the whole plan. Remote drafting means the person initiating the plan doesn’t have to wait for an idle terminal. And the choice between local and cloud execution means teams aren’t forced into a workflow that doesn’t fit their environment.
According to Mitch Ashley, VP and practice lead for software lifecycle engineering at The Futurum Group, “Ultraplan separates the planning phase from execution in AI-assisted development, turning terminal output into a cloud-based, reviewable artifact before any code changes. That structural shift addresses a real workflow gap for teams coordinating migrations or service refactoring.
The execution flexibility is consequential. Teams can approve a plan in the browser and execute it either in the cloud or the local environment. That choice is what makes distributing work between cloud and terminal sessions practical for higher-stakes development.”
The requirement for a GitHub repository and a Claude Code on the web account does limit who can use Ultraplan today. But for teams already running Claude Code in a GitHub-connected environment, the ramp is short.
One practical note: Ultraplan and Claude Code’s Remote Control feature can’t run simultaneously. Both use the claude.ai/code interface, so launching Ultraplan disconnects any active Remote Control session. It’s worth knowing before you start.
Still in Research Preview
Anthropic describes Ultraplan as a research preview, which means the feature is live but still evolving. Behavior and capabilities may change based on user feedback. That’s worth keeping in mind if you’re evaluating it for team use — it’s promising, but not yet a stable production feature.
For individual developers and teams willing to test early-stage tooling, though, the core concept is solid. Planning happens in the cloud. Review happens in the browser. Execution happens wherever it makes the most sense. And the terminal stays free the whole time.
That’s a workflow improvement worth watching.