Google has rolled out Jules, its smart AI coding assistant, into a global beta phase. Previously seen only in private preview, Jules now gives developers everywhere a fresh way to manage pull requests through AI. Originally launched alongside Gemini 2.0 last December, this tool now runs on the Gemini 2.5 Pro platform, offering five free tasks a day to get you started. It’s a solid option if you’re looking for an alternative to tools like GitHub Copilot or OpenAI’s Codex.
Instead of simply suggesting code completions, Jules takes a more thoughtful approach. It launches a disposable Cloud Virtual Machine, clones your repository, and outlines a multi-step plan before making any changes. Whether you need to update dependencies, refactor code, add documentation, write tests, or address open issues, Jules piles these tasks into a neat GitHub pull request for you to review.
What sets Jules apart is its deep understanding of codebases. Leveraging the latest multimodal Gemini model, it can sift through extensive file graphs and project histories while sticking to specific repository guidelines. With the beta phase eliminating the wait-list, you can simply sign in via GitHub at jules.google and start assigning tasks using an upcoming assign-to-jules label. There’s even a free daily quota to ease you into the process, with options for more robust usage and enterprise controls likely coming later this year.
Notably, Microsoft has already introduced a similar coding agent in GitHub Copilot, which handles bug fixes and feature implementations in a similar manner. However, Jules integrates the planning, diff generation, and pull request creation into one smooth workflow, which could prove particularly beneficial for teams already working with Google Cloud.
Jules is listed on Google’s Labs experiments page and has its own marketing site. You can start exploring its capabilities immediately or join a live AMA on May 21, 2025, at 22:30 UTC on the Labs Discord server. With this shift from basic code-completion to full-fledged agent-based development, Jules offers a low-risk opportunity for teams to experiment with AI-driven pull request filing. The real test will be whether the agent can extend its functionality beyond languages like Python and JavaScript while matching GitHub’s depth of repository context.
