JetBrains IntelliJ IDEA 2024.3 allows local language models in the AI assistant
The Java IDE offers new features for Kubernetes applications, a revised user experience and the selection of the AI chat model in AI Assistant.
(Image: Ground Picture/Shutterstock.com)
JetBrains has provided its development environment IntelliJ IDEA, which is primarily geared towards Java and Kotlin, with the third major release of the year, version 2024.3. Developers will find a revised Structure tool window, easier debugging of Kubernetes applications and the stable version of the K2 Kotlin mode.
The paid AI service JetBrains AI Assistant 2024.3 is also available for users of all JetBrains IDEs, allowing them to choose between various AI chat models, including local ones, in the new release. In addition, the JavaScript and TypeScript IDE WebStorm, the .NET IDE Rider and the Ruby-on-Rails IDE RubyMine as well as the Visual Studio extension ReSharper and the IntelliJ Scala plug-in have reached the new major version.
Videos by heise
IntelliJ IDEA: Updates for Kubernetes, Kotlin and UX
IntelliJ IDEA 2024.3 is designed to make debugging Kubernetes applications easier: The Add Tunnel for Remote Debug option can be used to make a workstation a virtual part of a Kubernetes cluster, enabling local debugging of microservices with the preferred tools. Another update in relation to Kubernetes is the cluster-wide Kubernetes log access including streaming and pattern matching. This creates a centralized view of all events across pods, nodes and services. Both features are reserved for Ultimate users.
There is also news for Kotlin developers: The K2 Kotlin mode, which entered the alpha phase in March 2024, is now stable. In this mode, IntelliJ IDEA uses Kotlin's K2 compiler for faster and more robust Kotlin code analysis. The K2 compiler reached stable release in Kotlin 2.0 in May 2024 after three years of development.
Ultimate users will notice another new feature when using the IDE: In the Structure tool window, you can now view the Logical code structure in addition to the familiar Physical structure. This shows not only classes, methods and fields, but also links and interactions between the components of a project.
(Image:Â JetBrains)
In addition to other updates for the user experience, the spelling and grammar checker is now also active during indexing and the default tab limit in the editor has been raised to 30. The IDE can also handle the Scala programming language. Support for experimental features of the language, including named tuples, has been added for this purpose.
AI Assistant 2024.3 allows local chat models via Ollama
Available in all JetBrains IDEs, the AI-supported AI Assistant 2024.3 service comes with a number of new features. According to the manufacturer, one of the highlights is the option to choose between the Google Gemini, OpenAI or local chat models.
(Image:Â JetBrains)
Gemini 1.5 Pro 002 and Flash 002 have been added to the available LLMS (Large Language Models) from Google. Developers can use a local model via Ollama. Using such a model in the chat requires activating Ollama support in the AI Assistant settings and configuring the connection to your own Ollama instance:
(Image:Â JetBrains)
For all major programming languages, the update has also improved code completion and revised context management and the generation of inline prompts within the editor.
Download for Windows, Linux and macOS
IntelliJ IDEA is available in two versions: the paid Ultimate Edition and the free Community Edition. The differences in functionality are listed on the JetBrains website. The new release for Windows, Linux and macOS is available for download there.
All further information on IntelliJ IDEA 2024.3 can be found on the JetBrains blog and the "What's New" page.
(mai)
