JetBrains has just rolled out a pretty significant update to its AI features, bringing some welcome changes for developers using its popular IDEs. This comes a week after the company announced that it is sunsetting Aqua, its dedicated IDE for test automation, citing lower than expected adoption rates. While Aqua is fading out, it seems JetBrains is doubling down on AI within its core product line.
Arguably, the biggest change is the introduction of a new free tier for JetBrains AI. Before this update, accessing most of the AI Assistant features that rely on cloud models required a paid subscription. Now, the company is making these AI capabilities more accessible to everyone. The revamped subscription model brings both the AI Assistant and Junie under a single umbrella plan covering Free, Pro, and Ultimate tiers. To clarify, JetBrains AI Assistant and Junie are two distinct tools. The AI Assistant is more of a reactive helper, providing suggestions, explanations, and code snippets on demand within your current context, whereas Junie is designed to be a more autonomous agent capable of planning and executing multi-step coding tasks across your project.
With the 2025.1 release, JetBrains IDE users can now access all JetBrains AI features for free. This free tier includes unlimited code completion and support for local AI models, allowing you to use some AI features without an internet connection via tools like Ollama or LM Studio. There is also a limited quota provided for features that rely on third-party cloud models, giving users a taste of the more powerful capabilities.
For those subscribed to the All Products Pack or the dotUltimate bundle, there is more good news. JetBrains AI Pro is now included in these subscriptions at no extra cost. The Pro tier gives you a larger quota for using features powered by external cloud models, such as the AI chat, the multi-file editing capability, generating code, and smart workflows for things like refactoring, creating unit tests, and getting commit message suggestions. JetBrains notes that if you already paid for AI Pro separately and are eligible for this bundle inclusion, you might be able to double your AI Pro quota or request a refund. There is also an AI Ultimate tier available as a separate purchase for those needing an even higher quota.
The AI Assistant has picked up several smart upgrades in this update. One standout improvement is smarter context awareness. It now has a deeper understanding of your entire project, using advanced retrieval-augmented generation (RAG) to locate the most relevant files, methods, and classes to guide its responses. It also factors in your recently accessed files, making interactions more in tune with your current work. On top of that, you now have more control over what the AI sees, with the ability to add or remove file attachments sent to it.
Building on that improved context, JetBrains has introduced a beta multi-file changes feature in edit mode. This allows the AI Assistant to suggest modifications that span across multiple files in your project, which could be a real time-saver for implementing features or refactorings that touch several parts of the codebase. The AI uses RAG here too, to figure out which files are most relevant to the change you are requesting.
Another workflow enhancement is the new "Apply" button for code snippets generated in the AI chat. Instead of manually copying and pasting, you can now hit Apply, and the IDE will help you automatically insert the suggested code snippet into the optimal spot in your code. This streamlines the process of incorporating AI-generated code.
JetBrains is also expanding the AI Assistant's ability to get information from outside your immediate project code. The Model Context Protocol (MCP), currently in beta, allows the AI Assistant to securely access external data sources and tools by connecting to MCP servers. This means you could potentially provide the assistant with context from third-party tools, file systems beyond the project, or APIs, giving it a broader base of knowledge to work with.
There is also now a /web command in the AI chat that lets the assistant fetch documentation, troubleshooting tips, and other resources directly from the web without you having to leave the IDE, eliminating the need to switch tabs for a quick search.
For those who prefer working offline or want to use local models, the update introduces an offline mode. This lets you work without an internet connection and still benefit from some AI capabilities like chat, code generation, commit messages, and inline documentation by using local models via tools like Ollama or LM Studio.
JetBrains has also expanded the selection of large language models you can choose from in the AI chat interface. You now have access to newer and more powerful models, including: Claude 3.7 Sonnet, Gemini 2.5 Pro, and OpenAI GPT-4.1.
Finally, adding a layer of privacy and control, the AI Assistant now supports the .aiignore file. Much like a .gitignore file, it lets you specify files or folders that JetBrains AI should not access or process. This helps ensure that sensitive information or irrelevant build artifacts stay out of the data the AI uses.
