JetBrains Air: A System of Products for Agentic Software Development | Hacker News Reader

JetBrains Air: A System of Products for Agentic Software Development

blog.jetbrains.com

75 pointsFinnLobsien122 comments

I like JetBrains as a company, their pricing for individuals made a lot of sense, you buy the license and you can keep using perpetual license for that version, if you want to buy newer version, you pay slightly less than year before.

I want them to succeed, but can't see how they can win in this agentic world, where people are barely opening their IDEs, and living in agents terminal whole day and from time to time opening a PR to get a review

Agree, and I have been an Ultimate Edition license holder since like 2013.

I have opened an IDE for very very few things in the last 9 months or so. Im still renewing a license at this point mostly because I like JetBrains the company and it's not a financial burden but they are in a bad place.

If their revenue-generating products are not doing so hot, I wonder what that means to projects such as Kotlin and Compose Multiplatform. Will they keep the lights on?

The whole "multiplatform" portability business has been one of the most conspicuous victims of the agent wave. Porting is now primarily a QA exercise, not a development one. (Development too, generally, is heading that way).

JetBrainer here. Thanks for the loyalty, and I do think as one of the last independents, we are worth cheering for. But it's up to us to earn double-clicking on the IDE icon.

After using Air in PyCharm, it's some good stuff. It's re-imagined for this hybrid world. But my question for you, honest question: do you think you still want to look at code? How about write code? If so, what are some things you'd do, if you were in charge of our IDEs?

same here, but this will probably be my last year, their AI strategy where it seems impossible to use it with local models without it "calling home" really soured things for me, if I wanted an always connected spying on me IDE I would use vs.code, I paid for jetbrains because I wanted to use all functionality (including AI, pointed to my llama.cpp) with networking blocked in little snitch save quick unblocks for the occasional license check if needed.

JetBrainer here. I use OpenCode and Pi connected to local models and to stuff like hosted DeepSeek all the time. Once you have your model config set up in the agent, it works as expected.

thanks for the reply, last time I tried this in goland if you blocked the IDE from networking via little snitch/opensnitch, it did not seem to work or keep working, there seemed to be some sort of call home to enable the AI plugin / keep AI working. If this has changed I will retry and see.

Note https://youtrack.jetbrains.com/projects/LLM/issues/LLM-17450... is still open although https://youtrack.jetbrains.com/issue/LLM-16479/Be-able-to-ac... appears to be closed as completed now

I use their MCP to give agents reliable refactor/rename, find usages, etc.

We not only are opening our IDEs, the alternative are iPaaS products like Workato, Boomi, Opal, Sitecore AI, Power Platform,... which are IDEs in their own way, the revenge of 4 GLs.

I'm realizing I am a minority, but my IDE is becoming MORE valuable to me, not less, with LLMs.

The volume of code is only increasing. Problems that used to be solved with a few lines (or even removing lines) are now solved with changes across 10 different files.

My IDE is invaluable for me when trying to navigate this stuff. I am completely convinced that people reviewing LLM code only via PR diffs are consistently missing massive systemic issues. You *need* to review changes in context of the surrounding code.

And doing this without an IDE is just making your life harder for no reason.

So you're using the IDE for review? The main features you need are around simple things like reading files, tracking git changes, things like that?

That's a small subset of the functionality you would need if you're actually authoring code. So the IDE maybe be more valuable, but the feature set you need is actually much smaller and can probably be redesigned given your narrow workflow.

Is this accurate?

Sure, but my IDE does these simple things very well.

Just some important things off the top of my head that I use daily:

* Instantly jumping between definitions and usages

* Folding and unfolding code blocks

* Quick overview of structures in files

* Database access, so I can browse the data and the code at the same time

Also, sometimes I just want to quickly refactor/rename some variable across all its usages or whatever. It's just quicker to do it in the IDE with a simple keyboard shortcut, rather than prompting an agent "rename oldName to newName".

JetBrainer here. I can attest that you are a primary target for Air for IDEs.

We have put a lot of thought into how we can add value for the workflow you just described. Here's the way I think about it: FSD for a car. Can you imagine trying to sell a car without a steering wheel? I still want to code, I still want the deterministic side, even in cases where I'm letting the machine do the work.

I wanted to try Air, as it seems to have interesting ideas for working collaboratively with LLMs.

But the choice of AI providers was very limited, and I found no ways to add the ones I use (the only options for providers were Anthropic, OpenAI and Google).

Any plans of opening up for more providers?

With ACP, it allows for any provider you choose.

Cool, I'll try this today.

I would really like Air to be closer to an agent / terminal multiplexer like CMUX / WMUX: I would for example like to be able to basically type "claude" or "codex" to launch my harness with the CLI args I need. The fact this requires dedicated API keys from the LLM providers makes it a no go at both home and work.

I'd agree with this. I've been working with Claude on an app. I wanted to increase the size of one element in the UI. I could have asked Claude to do it, but I wanted to understand the application, as well as it kinda seemed like a waste of time and tokens.

Likewise.

Unless you are blindly vibe coding everything without an IDE to help navigate what is being generated to review the output, working would be hellish.

And if you are blindly vibe coding everything, I am just glad I don't have to work with you. While LLMs mostly speed me up, they make serious mistakes often enough that a review is necessary.

I'm biased (I'm a dev advocate at JetBrains), but I also still actively write code and use PhpStorm almost daily.

Yes, we're writing less code, but the job of "a programmer" has always been about so much more than _writing_ code. It's about code reviews, QA, debugging, discovery, and maintenance; much more than the "writing" part. Anyone with production code will agree, I think: the writing part is the easy part. That's where I still very much prefer a proper IDE over a text editor or terminal. I want easy code navigation, I want proper static analysis without relying or waiting on third party tools, I want a visual overview of changes that happened. I want to stay in charge of whatever is writing my code. On top of that, I want my agents to have access to the tools that have made me the most productive developer for years.

Curious about your thoughts?

> much more than _writing_ code. It's about code reviews, QA, debugging, discovery, and maintenance;

Of course, but agents are also opening up new possibilities to do these jobs.

For example, you can do quick first code review with agent, or you can ask agent to add couple of print statements and debug the code until it figures out the issue, also for discovery, you can do discovery with agents and generate nice output with the details you want (although not accurate sometimes)

As a JetBrains user myself, whose organization cancelled the IntelliJ license recently, I can attest things I am doing inside IntelliJ is getting less and less. I am primarily using IntelliJ for code navigation at the moment, which is still the best one

> It's about code reviews, QA, debugging, discovery, and maintenance; much more than the "writing" part.

AI can do all of that quite well. If you take your hands off of the wheel, lean into it, start building systems for it, start figuring out where it goes wrong on your codebase and building systems to prevent that... you might get 10x more done. I do, and the results are generally better than anything I'd have coded by hand, because there was never enough time.

I don't look at or try to grok all changes, though. I couldn't get near as much done, otherwise, and I'm not convinced that 1 hr of me looking over it is better than 4 hrs of AI looking over it, and the AI is far cheaper. I look over the code, occasionally, to look for code that's objectively terrible, but I don't fix it directly - I use it as a test-case for improving my AI stack, to make it able to find and fix or prevent that class of terrible.

I probably spend 9 out of 10 credits on review and refinement, not writing fresh code. In spite of that, I'm still producing things way faster than I ever did by hand, with only a single $200/mo Codex plan.

I still keep my JetBrains' subscription, so that agents can use the MCP. I probably spend <5h per month in an IDE myself, though.

Hi, JetBrainer here with my two cents. You're right, there's a crop of folks barely opening their IDE, who don't really even want to look at code anymore, much less type any code.

But there's still a group that needs to own the result, who still read code and write code. Folks who still believe in developers and development. This group might not be loud in SV, but they are big out in the world.

I'm a Python advocate at JetBrains and our team runs into folks like this all the time.

I've been enjoying using Air inside the IDE as a way to get the acceleration without giving up coding. I get the speed side from the agent and the trust side from the deterministic IDE.

We just haven't had the IDE moment for agentic development yet.

I suspect we won't for several years yet - and it's likely going to be way more integrated/specialized for specific technology than our previous ides were.

But you're unlikely to go back once you've the full stack integrated with full observability, release pipeline, system integration tests etc pp

Still, it needs way more integration to work well and not be DOA, hence this is still far out I think.

JetBrainer here. Kind of my thoughts as well. When you think about the main problem -- human-in-the-loop -- the deterministic side has a lot to offer. But not always at the point of cursor-on-a-character.

I think this is it. It's hard to tell what the future of IDE development (if any) looks like at the pace AI is going. So everyone is building agentic workflows for a fast moving target. What will happen when the code generated is already 99.9% trustworthy and doesn't need the human-based tooling of an IDE anymore? It feels like we're almost there.

I do think JetBrains has an advantage here with their background and reputation. I'm constantly reshaping my tools literally every month, so hopefully they build some type of staple development IDE to fit these type of workflows. I don't know what that is yet.

AIR looks interesting, but I can't say this is the "AI IDE moment" yet.

I'm still really soured on JetBrains for how abysmally they handled their UI update and initial AI offerings. I was an Ultimate subscriber for over a decade and canceled last year because JetBrains decided they weren't interested in having people like me as a customer. That being a power user who values a certain level of usability and continuitiy in their tools. Fischer Price's My First IDE is not a tool I can or am willing to use professionally, and I don't see the value of a paid VSCode clone.

Not to mention all the lies about the new UI adoption and when faced with backlash, they put the old UI into a plugin that "might" receive updates and support in the future.

I actually did try their latest IDE a few months ago to see if their AI integration was worthwhile. There appeared to be no functional way to use a local LLM. You can set up OpenCode as a provider, but any time you click anything it resets your model to one of the free API services, with no discernable way to change the default.

I don't have high confidence in Air. So far JetBrains has not shown they know what a good AI integration is like. Frankly I don't think they know who their customers are or what they want.

For my part, I'm still using 2024.1 versions on my perpetual fallback license. I had Claude slap together an MCP plugin so now I have my own AI integrated IDE that actually works and functions as a usable tool. All my buttons have text labels instead of hieroglyphs and padding.

I’ll try this out but after over a decade as a paid JetBrains customer I’ve been looking at the door. I still love their IDE and moving around the code in it but the IDE uses way too many resources. Even with me excluding the worktrees folder it gets stuck indexing/processing when I spin up new worktrees to the point that I’ve found myself killing IDEA regularly to regain system resources. 99% of my work happens in the terminal now and JetBrains has been very slow to adapt. If I get one more “can’t checkout branch, it’s checked out in a worktree” error I’m going to lose my mind, give me an auto-heal option or something. This is how work is done, don’t make me leave your IDE to fix it myself on the cli.

I’ll try out Air but I don’t have high hopes, my current workflows work great and I’m still not sold on what Air offers even after reading this post and the page about Air.

I tend to feel fine with IDEs using buckets of resources, after all: they are specialist tools.

But I have been using their jetbrains remote system, and you can clearly see how incredibly high the RAM usage is on both server and client.

Personally I am also looking at the door, but not because of the resource usage… but because they keep putting resources into things which aren’t their core offering and it makes me wonder if they have run out of innovation in that area.

… then again, they did recently figure out how to make development on a Windows remote work… so I can keep using my linux machine while developing Windows software, so I think that earned a lot of goodwill from me.

I find I still use the paid IDE's (RustRover and WebStorm), even when using LLM's for reviewing code and applying corrections. It's often just faster to do that, especially with an IDE which allows for quick non linear navigation through the files (following functions etc.). I think this is where JetBrain's IDE's excel. I can live with the high startup cost as long as it is quick afterwards.

I do feel however they have been neglecting anything non-AI for a while now. I in particular loved their auto complete, and with the new guesswork AI auto complete it's just not the same. I want deterministic behaviour from my IDE.

On the LLM side, I couldn't use the Anthropic subscription with their offerings, so I just use it from a harness outside the IDE, and switch to the IDE for review. This may not be JB's fault, but it still makes their offering useless to me.

I may be one of the few remaining but I also still program myself sometimes (I weigh it based on the task). The IDE has been pushing so much AI it has been getting in my way sometimes, and it is definitely using more resources for stuff I don't use.

I too am on the fence on whether I'll resume my subscription. Luckily for JB I dearly dislike VSCode's code navigation and interface, and would severely miss their debugging tools integration.

> I couldn't use the Anthropic subscription with their offerings

We tried to make this work, but unfortunately hit a hard stop with Anthropic: their current policies do not allow us to support this integration :c

We still hope to resolve this in the future

JetBrainer here. For both this comment and the parent...I get you. It feels like the whole industry has gotten sucked into this vortex where people don't even look at code anymore.

I'm happy about our Air launch because we're being AI "realists"...neither too much sunshine pumping nor too much skeptic. We're doubling down on developers and development, both when you type and when you don't.

On to the point..I can promise the PyCharm is doing some pretty massive non-AI stuff on the deterministic side. I think determinism is the best path to trusting the code generated by the box.

IntelliJ was the perfect IDE ... two years ago maybe? Very little that has been added since improved my experience, a bunch of it gets in my way, and I'm sick of being prompted to try the latest and greatest AI integrations.

I don't think that there's a better alternative at the moment, and I get the need to pump out features to keep the revenue stream going, but I miss the slimmer versions

JetBrainer here. About worktrees, I've seen the internal demos and we are making progress on that. It's actually an interesting problem, conceptually, about the index.

About the terminal, honest question...is that because you are looking at code less and less? Writing code less and less?

Yes, it’s because I’m writing less code by hand. I mostly use IDEA now to navigate/review the code or do my own debugging/tracing. My time in the IDE has fallen drastically which is why the high ram/cpu usage when I’m not even “using” the ide are so frustrating, especially since I’ve excluded the nested worktrees parent folder (.claude/worktrees). I can handle IDEA being dumb when it comes to worktrees, I can’t handle it locking up after I’ve told it to ignore them.

I know harnesses (and harness harnesses) have a long way to go and lots of improvements even without models getting “smarter”, I’m rooting for JetBrains but so far IDEA has been more of hindrance in the LLM-era. The CLI TUI’s have become the most important tool I have and the GUI versions haven’t kept pace.

I’d love to live in IDEA instead of the CLI but so far I’ve found that to be too limiting and not as adaptable to our rapidly changing field. Maybe Air will help with that, we will see.

I want something like herdr in IDEA, a way to monitor agents, smart worktree integration, easily visualize changes and use the power of the InteliJ brains to assist in coding (lint/checks/etc).

Right now my IDE gets more in my way than assists me in shipping code.

Nevermind, Air is DOA: "Authentication with Claude Pro/Max/Team plans is not permitted by Anthropic’s Terms of Service" - Sorry, but unless you can use a subscription then I can't afford to use your tool. I understand the ToS, I wish it was more open, but the reality is that API pricing is way too high and a subscription is all that makes sense currently.

I’ve been a happy paying JetBrains customer for 10+ years. This year I decided not to renew. It’s just not worth it anymore. I don’t spend as much time as i used to in an IDE. And their AI offerings are not so great. I gave Junie a solid try, but it was always too far behind the competition. I gave Air a solid try, but it was full of bugs. I wish them the best, and I hope they can find a solid product again in this new era.

Is anyone using any of JetBrains AI offerings? I feel like they are burning piles of money on areas I never hear anyone talking about using. The only thing I ever used was their AI plugin, tried Junie once, but the rest just seems like noise I dont care about.

Rider was my IDE for years until they started pushing their AI tools while not letting you bring your own models. Then it seems they listened to the complaints and for a while I could use Codex but then the next update took it away. So I ran out of patience and switched to VS Code, and that was the end of JB and my dear Rider for me.

Hey, JetBrains team here:

What is the problem with using Codex in Rider, can you please share? It's generally available, with your own Codex subscription or via JetBrains AI.

PyCharm and RustRover. It is astonishing no other IDEs have offered direct competition. I would prefer a faster/lighter alternative to my tablet, but none exists.

edit: Sorry; misread. No, and I am very confused by their diaspora of AI offerings!

Imo vscod(ium) with pyrefly LSP is miles better than Pycharm today. Never used rust rover but rust-analyser LSP also worked great for me. Interesting which features of jetbrains IDEs do sell it for you

I don't have a good direct comparison to that, but broadly:

  - Natural treatment of multi-file projects
  - Introspection
  - Refactoring
  - Keyboard shortcuts, specific types of search etc
  - The build/run windows

Overall, it feels like JB treats the core units as languages and multi-file projects, while others, including VsCode, treat it as text files. Whenever I try something else, I look for these features, assume I'm doing something wrong, then come to an understanding that no one else does them to the same degree. For example, Zed has no concept of moving items. I don't know if VsCode does, but whenever I try that, it seems to have only superficial ability to introspect on function calls, data structures etc.

The only IDE that compares in my eyes is Visual Studio, which to outsiders might seem weird, but when they support something they do a pretty good job.

I use Rider and DataGrip regularly. I also use WriterCide,

DataGrip is seriously under-rated, IMO.

As far as I know, you don't need DataGrip if you already have Rider. The same functionality is already inside Rider.

I like a completely separate tool for managing databases.

I mean, I use DataGrip, but their AI specific tools is what I'm wondering about.

Ah. I don't rate how JetBrains have gone about it. It feels like things have moved around a lot, and it seems too complicated, with too many AI "thingies" sprinkled around without much explanation. And the inline, auto-suggested change can vary from "useful, please do that", to "will you stop flashing up all this shit everywhere I look!".

For a while I used Junie for small refactorings, and it generally does a good job, quite quickly. A couple of times I tried asking it for more complex things, and it didn't do well. I do find the UI to be a bit tricky, so it's quite possible I did something wrong or didn't add something to the context I was supposed to. Hopefully they sort out the "AI Chat" panel soon.

But now I use Codex/ChatGPT... and what it can do is frankly incredible. And it might just be me, but I do like having my AI chat UI completely separate from my IDE.

I agree on DataGrip and use its MCP weekly.

I used to use datagrip, but one of my first "vibe code" experiments was to create a datagrip replacement for myself ( https://github.com/keithn/sql ) that worked better for the kinds of things I did, a lightning fast tui, but then I stopped using that and now just use claude mostly.

I use Junie at work and I have to say I'm not amazed by it. Plus I cannot run it inside an Alpine Linux VM which is annoying as all the other agents I tried worked just fine.

They are fighting for survival, hence why they are all here promoting a product that no one really wants but vaguely seems interesting with their barely used HN accounts. Might take a while for them to realize they are largely irrelevant and don't control the key aspects of the tech which is a deal breaker for most (can't used claude paid plans). They need to find a better pivot and likely shed a whole boatload of employees.

I would love to see JetBrains win, I like them as a business, though unfortunately I haven't been using their products in a long time now.

Hello, JetBrainer here :) That means a lot, thanks for saying it. If you ever want to kick the tires again, I'd be happy to point you at what s changed, a lot has moved, especially on the AI side. No pressure either way, just glad you're still rooting for us.

I think I am sort of "aware" of what's out there, though maybe not to its full extent, but one feedback I would have is to try and do one thing, and just one thing, really really reaaaaaally well.

There's so much going on right now, and the push to PaaS is not performing as well in my opinion.

Take a look at how Herdr got viral, and how it exposes protocols that let others come and build with it. Multiplexers are a good bet, though there are plenty of other verticals out there too.

I would say try to chase another ReSharper moment for you guys.

Semi-OT, but does anyone here know what exactly happened at JetBrains?

Their IDEs have been nosediving for years now in a way usually reserved to "has been acquired by private equity". But they have not been acquired by private equity as far as I can tell.

I'm writing this from my laptop, where each keypress in a .md file in WebStorm makes the fan spin up (not hyperbole).

What happened, JetBrains?

I'm still a happy payer, but... They started to get out of tracks with the "new UI".

The OP post feels like completely marketing department, no skin in the game. I'm afraid they are becoming Borland.

Follow-up question:

What else can one use?

Excluding VSCode and all those AI-first IDEs I mean. JetBrains' git features have been excellent. I have not yet seen them be reproduced anywhere, but, eventually, they will kill that classic UI plugin. :(

Haven't you noticed how their nosediving started right around another significant event of the past few years?

Most development was done in Russia (disregard tales sold to the West about them supposedly being a Czech company, they were Czech mostly in the name, real work happened elsewhere). Exited Russia in the beginning of 2022, leaving many talented people behind: not everyone wanted or could leave because of families, obligations, personal beliefs.

Remembering where most of JB's money was coming from, can't really say they had a choice.

That’s true, but I don’t think it’s the whole story.

In my experience, another major factor is Sergey Dmitriev’s (the sole owner of the company) management style. He keeps a tight grip on major decisions, with a fairly “my way or the highway” approach. That favors managers who are comfortable with the arrangement and can drive away people who want more autonomy or a say in the company’s direction.

For years, when JetBrains faced relatively little competition, that structure worked well enough. As the market changed, though, it became harder for the company to adapt.

JetBrainer here. I can promise: not...private..equity. In my conference talks I like to brag that I get to work at the last independent: private, profitable, European. It's a cool gig.

The answer is more like the obvious one...our category is getting redefined, at hyper-speed, and all of the players are all over the map trying to sort this out.

I like what we're doing with JetBrains Air. We know what we bring to the table, who it's for, and we're doubling down on developers and development.

That said, CPU fan on Markdown is not cool. Want me to look through it with you?

> Want me to look through it with you?

Honestly not really.

I mean I am happy that there is some kind of response, but this is just damage control.

And, frankly, you do not need me to fix that. Of that I am certain.

I'm not running any weird setups or something out of the ordinary. If you actually test your software on a normal computer (think: random thinkpad with 8th gen intel cpu), running some random boring linux distribution (think: debian) you will run into these problems.

The bad experiences all so far have been smoke-test-level issues.

Or.. well. Product strategy. Not even bugs or performance issues at all.

But, yeah, you had to say something. I understand that, given your role as "developer advocate" (as per the HN account about section). This is your job after all.

But, honestly, sticking ones head in the sand and hoping that this passes might've been the strategically more wise move.

The problems your employer has cannot be fixed by one person taking their time and looking at an individual issue some other vocal guy on the web has. I think we both know that.

I'm sorry, genuinely.

Nailed it, and have had this response before. Summarized:

  - This is a real problem with JB IDEs
  - It is not specific to an individual's setup
  - It has existed for many years
  - Anyone at JB who actually is interested in diagnosing or fixing it does not need to extract details, support tickets, or logs from users, as it's trivial to reproduce on any system.

Hi! Like many others here, I was a long-time subscriber of the whole suite, but hitting ten different bugs in Rider in a single day broke the camel's back - especially after I'd already given up on WebStorm once it became completely unusable on a medium-sized React codebase.

I really hope that you guys manage to turn the ship around, but after experiencing the decline myself and paying until this year, all my goodwill towards Jetbrains has been used up. At this point, I would never consider paying full price again.

Have you guys discussed (if you're able to share) allowing customers to resume subscriptions, which they've cancelled during the decline years? For me, that would be an important step to rebuilding trust.

Jetbrains ides were always slow and buggy. But with best in class code understaning

I may have observed this regarding performance. (RR + Pycharm) Perhaps the classic software move of: "[My] computers have gotten faster, but JB IDEs are still slow/unresponsive, so they are probably getting slower / more resource intensive".

They will bring a 9950x with nothing else running to a crawl. Then complain they are out of memory. (When it is obviously the IDEs themselves running the system out of memory)

I <3 JB IDEs for their functionality, but resource pigs they are.

/throwaway for obvious reasons/

The previous CEO has lost the business and wasn’t able to deliver any new products. He was replaced together with executive suite few years ago. As for me they are now even more clueless than before and don’t have any strategy, and very few of us are still excited about JB future.

I've been an Ultimate subscriber for many years. I started out using ReSharper in Visual Studio, then picked up ReSharper C++ when I began writing C++, and eventually moved to Rider. As I also work with Java and Python, I just went ahead and bought Ultimate. Even though my company provides Visual Studio and other tools, JetBrains has given me a genuinely enjoyable development experience all these years.

These days everyone is coding with Agents, but I still really enjoy writing code by hand. I like implementing features myself and designing architecture. But spending a significant amount of money every year on an Ultimate subscription just to support that hobby no longer makes sense from an ROI perspective. At this point, most of what I'm paying is really just to support JetBrains and to thank you for the enjoyment you've brought me over the years.

My subscription is about to renew again, and I'm still on the fence about whether to continue. It looks likely that I will, but if the ROI keeps declining, I'm not sure what the future holds. It's not that JetBrains' IDEs have changed, it's that the times have changed.

(JetBrainer here.) I get what you're saying, and thanks for supporting us.

I will say though: I hope we keep you and earn keeping you, because you are EXACTLY our bulls-eye. Folks who still believe in code and coding and still want to put their hands on the wheel sometimes.

people need to stop trying to build the same walled gardens with these development environments. nobody has time to buy in at this point or try to grok what your specific vision of "agentic development" is.

I have Ghostty with Claude open with one tab per project. Then I have IntelliJ open also with one tab per project.

I find this setup works great for actually participating in the development. To understand what is happening and keep the mental model of what the project is about. Review in progress work in IntelliJ, use the Git, Database and REST(.http files) in the IDE as the agents generate the code.

I don't see a problem with this setup is it really that uncommon? Have people stopped using TOOLS completely? That said I don't have much interest in these JetBrains AI products, at least not currently.

Similar workflow myself - claude cli + intellij project per worktree. Works well, intellij IDE tools are still unmatched compared to using Zed / VSCode / etc.

The "Air Alpha" plugin isn't bad to move the claude cli into the IDE window itself - but not essential.

Fun fact: IntelliJ itself switches to Ghostty VT (with Swing rendering)

https://github.com/JetBrains/intellij-community/blob/55910d2...

As the relevant issue states, it's exactly because it should perform better for TUI agents.

https://youtrack.jetbrains.com/projects/IJPL/issues/IJPL-226...

I've been paying customer for 15+ years, I wish I can continue using them another 15 years, but the current trend is terrible.

They had the best IDE. Now they went all into AI, trying to compete on another people's field, and the cost of that is diminishing their main product. Do they really want to compete with Anthropic? Instead of that they could improve the other sides of IDE. Even with AI I have to review the code. Especially with AI. I need tools to check the diffs, I need to refactor it, I need to run it in different ways, see how it works, and so on. That's where IntelliJ could shine, not as AI tool.

IntelliJ still provides the features you've listed. However, the developer workflows have changed quite a bit, and the tools have to adjust to that.

Right, and those features are deteriorating though it should be opposite as it becames more important now

The problem all these products have is these blog posts that describe in enough detail what the patterns are, out of necessity, also provide the information for the product to be cloned into a form that is then more dynamic and malleable for use by other developers.

It's clearly existential for JetBrains, and it's revealing that the opening goes for the fear with "but it hasn’t changed what it costs to be wrong" while ignoring that this hasn't exactly been a major concern for most software people for a scarily long time, including JetBrains themselves.

I stopped using them last year. They were too late to see the AI juggernaut coming and their attempts at responding have been underwhelming. They should have been pouring resources into Fleet from 2022 onwards to make this, but instead it's been a series of missteps. This looks like approximately the right offering, but where's the moat? It seems like just another Cursor/Antigravity.

It is a broader story than just the IDE integration. That's important, of course, but the post talks about agentic development across the IDE, teams, and organizations.

not to mention it's super easy to build your own tooling

They need to fix how their AI Chat works. I could not for the life of me get it to edit files when using a local model with LM studio. It gave me the correct code to change. There is no Code mode when connecting to LM studio with a local model.

I couldn't find how to change the mode without it trying to use Junie, or claude etc. very frustrating.

Hi, on this, I can promise (as a daily user) that our OpenCode and Pi support are really good. And with those, it's really easy to connect to LM Studio...in fact, I did this for the DeepLearning.AI course on hybrid/local models.

> Authentication with Claude Pro/Max/Team plans is not permitted by Anthropic’s Terms of Service

Falls at the first hurdle...

Isn’t that a Anthropic limitation? Their TOS only permits their own harness?

Well I've no idea what this app is, because I can't use it, but it sounded like somewhere that I could run claude code inside, guess I was wrong.

I am unlikely to renew my ReSharper licence without at least a 75% discount:

- It sees very little investment ever since they introduced Rider.

- They give out VS Code one for free, yet the VS users are neglected.

- Touching code manually is outdated by agents anyway.

Refactoring when you know the shortcuts and what you want to do is still significantly faster and less error prone than letting an llm do it. That is why I started semedit[0], to bring semantically editing to LLMs. I’m still working on the benchmarking harness but the first results where I get the llm to actually use the tools look promising.

[0]: https://github.com/spockz/semantic-editor

We are sorry to make you feel like we de-prioritized ReSharper. Actually, we do not. For quite a while we were busing with working on one of the most important change in ReSharper - making it work out of VS process. This is now a default mode starting with 2026.2.1: https://www.jetbrains.com/resharper/whatsnew/#out-of-process This was also accompanied with a huge work on various performance issues and freezes.

Another big thing for ReSharper was unlocking VS users to use more agents in VS: https://www.jetbrains.com/resharper/whatsnew/#preview-of-acp.... Next release we plan to bring in Claude Code and OpenAI Codex via this ACP entrance. And then we will combine that with the classical ReSharper intelligence in the format of skills, hooks, etc.

As for the free tier, in VS Code it's indeed there for non-commercial usage, but we don't see that many non-commercial users in VS. Majority of ReSharper users are organizational clients. That's why our decision on licensing. Hope that explains. But open to any feedbacks here.

I think it makes total sense for JetBrains, and a bet on multi-vendor is something that speaks sanity to me. JetBrains has never talked VC bull**t and I'm glad they are not losing their minds for AI

(JetBrainer here.) When I give conference talks, I introduce JetBrains as: private (never taken VC), profitable, and European. We are indeed focused on being AI Realists, on the side of coders who still want to code.

Highly recommend SoloTerm, which I think this kind of tries to replicate.

I still pay for it but it’s getting harder to justify spending on it.

They have to rapidly pivot into AI because their IDEs will soon no longer be needed because less and less people are writing code by hand but I don't think it will be successful.

Not sure what Jetbrains can offer anymore. I haven't used their tools for 8months or so, looking to cancelling all products pack before next renewal. Rider was great for years, but last couple of years it's just become brutally slow. Now everything can be done in a terminal and an editor like gvim (or any editor really). Just get claude to open code in your editor when you need or even create visual code reviews, or any number of ways at looking at code with claude. All the refactoring tools jetbrains had are out the window, debugging with a debugger is largely dead, air seems super meh and not being able to leverage a plans allowances makes it a non starter. The one thing we will keep is youtrack, it works well with AI. But weirdly replaceable if they do something stupid with it / price gouge etc.

I think this is the future of code editors.

Two things that fell short for me:

1. No support for my Claude subscription

2. The git interface lags behind other JetBrains IDEs

Hi, product person from JetBrains here:

About point 1: We tried to make this work, but unfortunately hit a hard stop with Anthropic: their current policies do not allow us to support this integration, and we were not able to negotiate an exception :c

We still hope that we will find an agreement here

That's unfortunate. I hope you guys are able to crack a deal.

Also, is it possible to use JetBrains Air with OpenAI-compatible APIs/proxies?

JetBrainer here. Yes, you can run your own local OpenAI-compatible models in Air, including through Ollama or LM Studio, or via open source agentic frameworks like OpenCode or pi. We offer a huge number of agentic frameworks that you can easily install to integrate with the IDE via the ACP.

I mean the OpenAI API integration but on a different hostname, like another URL that is OpenAI compatible, not with another ACP following agent framework.

Try this in Zed:

Open settings > AI > LLM Providers [Configure] >

there is OpenAI and OpenAI subscription - ignore those two.

On top-right, there is "Add Provider" with two options:

1. Compatible APIs (OpenAI)

2. Compatible APIs (Anthropic)

When I choose "Compatible APIs (OpenAI)", I can input a custom API URL.

OK I checked and it is available in the Plugin version of Air, but not in its own IDE. Which is okay.

Any ETA on the mobile app?

No integration to Claude subscriptions seems like it probably makes dead in the water which is a shame

Hello, As my colleagues already replied earlier to someone else here, we tried to make this work, but unfortunately hit a hard stop with Anthropic: their current policies do not allow us to support this integration, and we were not able to negotiate an exception

How come I can do this in Zed?

You can actually plug Zed's ACP into Air's acp.json file to draw subscription tokens:

{ "agent_servers": { "Claude Code": { "command": "bunx", "args": ["@zed-industries/claude-code-acp"], "env": {} } } }

Swap bunx with npx -y if necessary

I want to like this, but not being able to plug this into eg. OpenRouter or a local agent is a bummer.

Maybe I'm misunderstanding, but isn't a sizeable part of the article about how Air can be used with any agent (section "The future is multi-vendor")

It supports ACP, so you can use it with any ACP speaking harness including many that would work with local agents

Yes, as others replied, just plug in an agent with a model config pointed at OpenRouter. I do this all the time with OpenCode and Pi. Give me a reply if I can help you on this.

OpenCode fixed it for me, I just asked it to set up itself as an ACP server for Jetbrains Air. Running with that now.

Hadn't used ACP yet.

This is like a 3rd or fourth Jetbrains product named Air

JetBrains is on track to become another Nokia.

Just yesterday tried Intellij ai assistant setup and it sucks. I mainly use Cursor but also have Codex sub, and since version 3.15 Cursor made changes that nerfed Codex experience inside IDE, so I reluctantly look for replacement that could allow me to use both from a single interface. Intellij kinda allows for that unified interface, but is very prototype-y quality, missing features from both Cursor and Codex, can not import historical conversations and their auth UX is meh. Not sure if they will be able to catch up to competition from there.

hey darkwi11ow, would love to hear more about the fricition you experienced in the setup! don't want people immediately moving on when they haven't even had the opportunity to experience the tool!

Hey few things that are deal breakers for me (I tried Intellij IDEA community edition, AI assistant)

1) Import of historical data from agents. I often look up past agent sessions and require search into them as a primary feature of the harness. I know for a fact that historical conversations are not encrypted for both Codex and Cursor on a local filesystem, so Intellij tool could import them into ai assistance plugin even before user is authenticated with either Cursor or Codex. There are few annoying things in search for both Cursor and Codex extensions, so having great search over past conversations could be a great introductory feature

2) Auth .. I am not sure about this one technically because I haven't looked myself into how agent auth is done for Cursor and Codex, but at least for Codex, their IDE extension for vscodium/cursor/zed/etc. is able to use existing authentication from the CLI codex tool, so I do not need to login into every new IDE I try Codex extension in. With Intellij Codex auth did not work at all for me.

3) Cursor authentication worked, though I needed second one for Intellij while being logged in with CLI, but the model reasoning level is not selectable from within ai assistant, one more deal breaker for me unfortunately (I found it stays like this for a long time: https://forum.cursor.com/t/intellij-cursor-acp-reasoning-lev...)

I wasn't able to integrate Air into my workflow on my first attempt. I'd hoped it would unify Claude and OpenAI, but Claude doesn't let you use it unless you are on the Token plan, which IIRC only makes sense for enterprises. I did not see any immediate advantages in Air over the Codex application, but will try it again in a few months.

I still love RustRover and PyCharm! Switching to anything else would be a big downgrade.

Side note: I am confused by the various JB LLM integrations. Air, Junie, at least 1 (maybe 2 more?) the IDEas have suggested I try.

There's an ACP connector that allows you to use Claude with a subscribtion.

It's not bundled with the IDEs but it's available on github

I could def go for some ACP RN.

{ "agent_servers": { "Claude Code": { "command": "bunx", "args": ["@zed-industries/claude-code-acp"], "env": {} } } }

is to be put into your Air's acp.json (or swap bunx with npx / pnpm exec if you have only one of them)

JetBrainer here. You're right about the Claude terms of use issue.

About compared with Codex, we have some tricks up our sleeve that are actually valuable for the target market we're going after: folks that want to understand and own what they've done. Watch for the videos we'll be doing (and look for my shining face in the PyCharm channel.)

On the last point...yes, the downside of choice is choice.

LFT the vids! Is there a good way to break down the overlap etc between:?

  - Junie
  - Air
  - AI autocomplete
  - AI Assistant
  - AI Chat

Reply on news.ycombinator.com