This article is one chapter of my book Delphi in all its glory – AI-assisted development, the fifth book of the series.
In the newer versions of Claude, a tip that pops up in the terminal says: “Run tasks in the cloud while you keep coding locally.” At first glance, this sounds like marketing fluff. After all, if Claude Code is busy in one terminal, you can just open a second one. Problem solved, right?
Not exactly. There are a few real differences that make the cloud version worth knowing about.
What Is It?
Claude Code can run not only in your local terminal but also on Anthropic’s servers, through the web interface at claude.ai. You go to claude.ai, give it access to your GitHub repository, describe a task, and walk away. Claude clones your repo in the cloud, works on it autonomously, and when it is done, it pushes the changes to a new branch. You review them, and one click turns them into a pull request.
The whole thing runs on their infrastructure. Your computer is not involved at all after you hit Enter.
Why Not Just Open a Second Terminal?
That was my first reaction too. But there are two practical differences that make the cloud version more than just a fancy second terminal:
You Can Close Your Laptop
When you run Claude Code locally, the session dies the moment your machine goes to sleep or loses power. The cloud task keeps running on Anthropic’s servers regardless. You can fire off a big refactoring job on Friday afternoon, shut down your workstation, and come back Monday to find the finished branch waiting for your review. Try that with a local terminal.
No File Conflicts
This is the one that matters most for Delphi developers. If you run two local Claude Code sessions in parallel, both are working on the same files in the same project folder. One session edits FormMain.pas, then the other session overwrites the same file moments later. You end up with a mess. The cloud version avoids this entirely because it works on its own separate clone of your repository, on a different machine. It creates a new branch and pushes its changes there. Your local working copy stays untouched.
The Catch
There is always a catch. The cloud version wants your code on GitHub: it clones from there and pushes its branch back there. Without GitHub, claude –cloud can upload a local git repository instead, but then Claude cannot push its changes back. For many solo Delphi developers who keep their code on local drives or private servers, this is still a deal-breaker. You would have to push your code to GitHub first, which not everyone is comfortable with (even with private repositories).
Also, the cloud environment cannot compile your Delphi project. It runs on Linux, so there is no Delphi compiler available. Claude can read, analyze, and edit your code, but it cannot build it or run your tests. You will still need to do that locally after merging the pull request.
Finally, the cloud version sees only what is committed to your repository. The project’s CLAUDE.md, hooks, MCP servers and skills come along with the clone. Your personal ones, kept in your user profile, stay on your machine. The local version, with your personal configuration and your Windows tools, remains the more powerful option for day-to-day work.
This was one chapter of a whole book
You just read one chapter of Delphi in all its glory – AI-assisted development: more than 400 pages about Claude Code and AI for Delphi programmers. Written by a Delphi programmer, with the failures documented next to the wins.