Google today took Jules, its asynchronous AI coding agent powered by Gemini 2.5 Pro, out of public beta and made it generally available. The tool clones a GitHub repository into a Google Cloud virtual machine, uses Gemini 2.5 Pro to develop coding plans and fix or update code, and opens a pull request — all while the developer works on other tasks or steps away entirely.

During the beta, thousands of developers completed tens of thousands of tasks, resulting in over 140,000 publicly shared code improvements. Feedback led to new features including reusable setups for faster task execution, GitHub issues integration, and multimodal support. Google also updated its privacy policy to clarify that data from private repositories is not used for training.

Jules now has structured tiers: a free introductory plan capped at 15 individual daily tasks and three concurrent ones, down from 60 during beta. Paid tiers bundle into Google AI Pro ($19.99/month) and Ultra ($124.99/month), offering 5× and 20× higher limits respectively. The tool logged 2.28 million visits worldwide during beta, with 45% from mobile devices. Kathy Korevec, director of product at Google Labs, told TechCrunch that Jules is already used internally at Google for some projects, and there is a push to expand its use. The asynchronous approach positions Jules differently from synchronous coding tools like Cursor and Lovable, letting developers offload tasks and return hours later to completed work.