Google JULES coding agent now outside Beta

On Wednesday, Google launched an artificial intelligence coding agent, JULES, from Beta, after a little more than two months General examination for the first time in May.

Silled from Guemini 2.5 Pro, JULES is an unlimited coding tool that is integrated with GitHub and Clones Codes in Google Cloud Virtual devices, and artificial intelligence is used to repair or update software instructions with a focus on other tasks.

Google Initially JULES as the Google Labs project in December and made it Available for beta laboratory Through a general inspection at its I/O developer conference.

Cathy Corific, the director of the Google Labs producer, told TECHCRUNCH that the stability of the improved tool led to its removal from the experimental version after receiving hundreds of user interface and quality updates during the experimental version.

She said: “The course of the place where we will give us a lot of confidence is that Gul is present and will be for a long time.”

With the start of the broader operation, Google has provided the levels of the JULES organized pricing, starting with a free “introductory arrival” plan at 15 individual and three synchronous daily missions, below the maximum tasks 60 during the house. JULES paid levels are part of the Google Ai Pro and Ultra plans, which are $ 19.99 and $ 124.99 a month, and two subscribers offer 5 x and 20 x higher respectively.

Korevec noted that the Gul packages and their pricing are based on the visions of “real use” that have been collected during the past two months.

She said: “We have helped the maximum mission 60-Tagn to study how developers use Jul and provided us with the information we need to design the new package.” “15/day is designed to give people a feeling if Jul will work with them in the real project tasks.”

Google has also updated the JULES privacy policy to be more clear on how to train AI. If the warehouse is general, its data may be used for training, but if it is special, Korevec said it has not been sent data.

“We got a few comments from users as (privacy policy) were not clear as we thought, so we only respond to that. We have not changed anything about what we are doing on the training side, but we changed the language,” Correvik said.

During the experimental version, Google said that thousands of developers dealt with tens of thousands of tasks, which led to the participation of more than 140,000 improvements in the law publicly. Initial comments have led Google Labs to add new potentials, including reusing previous settings to carry out tasks faster, integrate with GitHub problems, and support multimedia inputs.

Corrivic said that the main users of Julus so far are artificial intelligence lovers and professional developers.

By non -simultaneous operation in a virtual device, JULES stands regardless of the highest AI coding tools such as the indicator, ligarforf, and beloved, which all work simultaneously and require users to see the output after each router.

“JULES works like an additional set of hands … You can start tasks mainly, after which you can close your computer and stay away if you want, then come back hours. Gul will complete those tasks for you, opposite if you do it with a local agent or use a simultaneous agent, you will be restricted to that session.”

This week, Gul got deeper integration with GitHub Open withdrawal requests automatically Just as it can open the branches – and a feature called Environmental footage To save the dependencies and install text programs as a piece to carry out the task faster and more consistently.

From the coding of Vepi to the use of mobile devices, the experiments of Beta Jules Development reported

Since entering the general experimental version, Gouls has registered 2.28 million visits worldwide, 45 % of them from mobile devices, for all data from the market intelligence provider similarReview by techcrunch. India was the highest traffic market, followed by the United States and Vietnam.

Google did not share details at the JULES user base and the best geographical areas.

Korevec Techcrunch told that during the experimental version, the team noted that many people used JULES from traditional coding tools to repair errors that may have been implemented or extended the promotional project to make it more willing to produce.

Originally, JULES requires users to obtain an existing code base. But Google soon realized that many potential users – like those trying to artificial intelligence tools – may want to explore it without one. Correvik said that the company soon enabled Gul to work even with an empty warehouse. This helped increase its range and use.

Google Labs also noticed an increasing number of users who reach JULES through their portable devices. Although the tool does not have a mobile phone application, Korevec said that users were connected to it through its web application.

“Since the great use that we see is emerging, we are completely exploring what features that people need a mobile phone a lot,” I noted.

Besides the beta test, Korevec stated that Google is already using JULES to help develop some projects internally, and there is now a “big boost” to use the tool in “more projects” in the company.