Google is further integrating its AI coding assistant Jules into developer processes by launching a command-line interface and opening up its public API. This enables Jules to connect with terminals, CI/CD pipelines, and platforms like Slack, as tech giants compete to shape the future of AI-powered software development.
Previously, developers could only access Jules—Google’s asynchronous coding agent—through its website or GitHub. On Thursday, Google rolled out Jules Tools, a command-line interface that embeds Jules directly into the terminal. With this CLI, developers can interact with Jules via commands, making it easier to manage tasks without toggling between the web and GitHub, and allowing them to delegate coding work and review results without leaving their usual environment.
“Our goal is to minimize the need for developers to switch contexts,” explained Kathy Korevec, director of product at Google Labs, during an interview.
Google already provides Gemini CLI, an AI-driven command-line utility that operates across developer tools like terminals and CI/CD workflows. Both Gemini CLI and Jules are powered by Google’s Gemini 2.5 Pro AI model. However, Korevec told TechCrunch that Jules Tools is tailored for “highly specific tasks,” while Gemini CLI is intended for more iterative and collaborative use.

Denise Kwan, a senior developer advocate at Google, further discussed in a Medium article how Jules differs from Gemini CLI. She pointed out that Jules is intentionally less interactive, carrying out tasks on its own once the user has approved its plan.
Alongside the CLI, Google has also released Jules’ API to the public, which was previously only used internally. Korevec said this move is intended to help developers incorporate Jules more seamlessly into their established workflows, where they are already comfortable and efficient.
Through the API, developers can also connect Jules to their integrated development environments (IDEs), such as VS Code, which provide a suite of coding tools. Korevec mentioned to TechCrunch that her team is also interested in developing dedicated IDE plug-ins to further extend Jules’ reach.
These new features follow Google’s recent addition of “memory” to Jules, enabling it to remember user interactions, preferences, suggestions, and corrections. In recent weeks, Jules has also gained several enhancements, such as a stacked diff viewer, image upload functionality, and the ability to read and reply to comments on pull requests.
Google is now also working to make Jules less dependent on GitHub. At present, Jules operates within a GitHub repository, requiring users to either link an existing repo or provide a new one for it to use.
“Users are asking for Jules to work with other code hosting services,” Korevec said. “We’re exploring ways to support additional version control systems, as well as options for those who don’t use version control or aren’t concerned about where their code is stored.”
Managing AI tools remains a challenge, particularly in professional environments. Jules, however, is built to alert users if it encounters a problem, prompting them to intervene if necessary.
“If it hits a snag or finds itself unable to proceed, it will stop and ask me for input,” Korevec explained.
Still, monitoring becomes more complicated when users access Jules on mobile devices, since native notifications aren’t yet available. Korevec noted that many people already use Jules via its mobile web version, and said Google is working to enhance the mobile experience, especially by adding support for native notifications.
To date, Jules has mainly been adopted by software engineers and other technical professionals, unlike many vibe coding platforms that target non-developers. Nonetheless, some users are trying out Jules as a supplement to more informal or creative coding tools.
“We often see users bring projects to Jules after reaching the limits of their current vibe coding tool, looking to expand them further,” Korevec told TechCrunch.
Jules entered public preview in May, left beta in August, and is now offered with tiered pricing. The free plan allows up to 15 individual tasks per day and three tasks at once. Higher tiers—Google AI Pro and Ultra—are priced at $19.99 and $124.99 per month, offering roughly five and twenty times the limits, respectively.