Google officially launched its AI coding agent Jules on Wednesday, marking the conclusion of its beta phase just two months after its public preview in May.
Powered by Gemini 2.5 Pro, this asynchronous agent-based coding tool integrates with GitHub to clone code repositories into Google Cloud virtual machines. It employs AI to fix or update code automatically, enabling developers to focus on other tasks.
Initially announced as a Google Labs project in December, Jules was made available for testing during the I/O developer conference. Kathy Korevec, director of product for Google Labs, told TechCrunch that after hundreds of UI and quality updates during testing, the tool's stability significantly improved, prompting its official release.
"The trajectory we're on has given us confidence in Jules' long-term potential," she stated.
With broader adoption, Google introduced structured pricing tiers for Jules, starting with a free "Introductory Access" plan allowing 15 daily personal tasks and 3 concurrent tasks—lower than the 60-task limit during beta. Paid tiers are included in the Google AI Pro ($19.99/month) and Ultra ($124.99/month) plans, offering 5x and 20x higher limits respectively.
Korevec noted these pricing models were developed based on "real-world usage insights" collected over recent months.
"The 60-task limit helped us study how developers use Jules and provided the information needed to design new pricing tiers," she explained. "The daily 15-task limit is designed to help users determine if Jules can work for their actual project needs."
Google also updated Jules' privacy policy with clearer explanations about AI training processes. Public code repositories may be used for training, but private repositories won't have any data transmitted, Korevec confirmed.
"We received feedback from users that the privacy policy wasn't as clear as we anticipated, so we've primarily focused on addressing this concern. Our training practices remain unchanged, but we've revised the language," Korevec said.
During testing, Google reported tens of thousands of tasks completed by thousands of developers, resulting in over 140,000 publicly shared code improvements. Initial feedback drove the addition of features like task setting reuse for faster execution, GitHub issue integration, and multi-modal input support.
Two primary user groups have emerged for Jules: AI enthusiasts and professional developers, according to Korevec.
Operating asynchronously in virtual machines, Jules differs from top AI coding tools like Cursor, Windsurf, and Lovable that require synchronous interaction. Users must review outputs after each prompt with these competing tools.
"Jules is like an extra pair of hands... you can basically delegate tasks to it and close your laptop if you want. You can return hours later to find the tasks completed. With local or synchronous agents, you're tied to that session," Korevec explained.
This week, Jules introduced deeper GitHub integration to automatically open pull requests alongside existing branch creation capabilities. A new "environment snapshot" feature saves dependencies and installation scripts as snapshots for faster, more consistent task execution.
From ambient coding to mobile usage, beta testing shaped Jules' evolution
SimilarWeb data reviewed by TechCrunch shows 2.28 million global visits since Jules entered public beta, with 45% from mobile devices. India leads in traffic volume, followed by the U.S. and Vietnam.
Google hasn't disclosed specific details about Jules' user demographics or geographic distribution.
Originally requiring existing code repositories, Google quickly adapted to user demands after realizing many potential users wanted to explore the tool without pre-existing codebases. Korevec said this adjustment expanded Jules' accessibility and usage scope.
Google Labs teams also observed increasing mobile access to Jules despite lacking a dedicated mobile app. Users currently access it through web applications.
"Since we see mobile use emerging as a significant use case, we're definitely exploring what mobile users need," Korevec noted.
Beyond external testing, Korevec revealed Google has already adopted Jules internally for project development and is now "aggressively pushing" to expand its use across "more projects."
