1
1
Published: November 13, 2026
Jules and Claude Code are both AI coding agents, but they work in very different ways. Jules is Google’s asynchronous agent: you hand it a task on a GitHub repository, it works in a cloud environment and returns a pull request. Claude Code is Anthropic’s interactive agent that you work with in real time from your terminal or editor. If you want to delegate small tasks and review the result later, Jules fits. If you want to collaborate step by step on complex changes, Claude Code is the closer match.
Everything else follows from this one distinction. With Jules, you describe a task, it clones your repo into an isolated cloud virtual machine, writes a plan, makes changes and opens a pull request for you to review. You do not pair-program with it; you submit work and come back to a result. This description matches how Morph’s March 2026 explainer outlines Jules, and that source also notes Jules works through GitHub, a web interface and a command-line tool rather than inside an editor.
Claude Code runs where you work, in a terminal or supported code editor, and you steer it through a live session: asking, reviewing, correcting and continuing. That makes it a better fit for exploratory or tightly coupled changes where you want to approve each step.
| Jules | Claude Code | |
|---|---|---|
| Made by | Google Labs | Anthropic |
| Working style | Asynchronous: submit a task, review a pull request | Interactive: work alongside it in a live session |
| Where it runs | Isolated cloud VM, connected to GitHub | Your terminal and supported editors |
| Free access | 15 tasks per rolling 24 hours, 3 at once | Not included on the Free Claude plan |
| Paid access | Included with Google AI plans (100 tasks a day on Pro, 300 on Ultra) | Included with paid Claude plans, sharing your usage allowance |
Jules does not have a standalone subscription. According to Google’s Jules usage page, the free plan allows 15 tasks per rolling 24 hours with 3 running at once. Jules in Pro is included as a benefit of Google AI Pro, and Jules in Ultra comes with Google AI Ultra, with Google’s coverage listing 100 and 300 daily tasks respectively. 9to5Google lists Google AI Pro at $19.99 a month in the US, though Google updates these plans regularly.
Claude Code, by contrast, comes with a paid Claude plan and draws from the same usage pool as normal chat. Our guide to the Claude Code subscription price breaks down Pro, Max and Team costs and how the limits work.
Jules suits well-defined jobs you keep postponing: fixing a small bug, updating a dependency, adding tests or tidying documentation. You can queue several tasks, do other work and review the pull requests when you are ready. Because every task is independent, it works best when each job stands on its own.
Claude Code makes more sense for larger changes where direction matters, such as refactors that touch many files, debugging an unfamiliar codebase or building a feature step by step. You see what it is doing as it happens and can redirect it immediately.
The tools solve different problems, so they are not strict rivals. Some developers will use an async agent for backlog tasks and an interactive agent for focused work. If you try both, start with the free options and compare how much review each pull request or session requires.
Pick Jules for hands-off, GitHub-based tasks you can review later, and Claude Code for hands-on collaboration on complex work. Neither is a replacement for human code review. For more comparisons of AI coding and productivity tools, visit our AI Tools & Apps section.