October 2026 Summaries
58 posts from JetBrains
Filter
Month:
Year:
Post Summaries
Back to Blog
Kotlin’s latest roundup highlights the Shipaton 2026 results, where BoatBrawl, AfterFade, and Lumeo won the Ship Kotlin Everywhere awards from nearly 4,000 submissions, demonstrating varied uses of Kotlin Multiplatform and Compose Multiplatform. It also announces Kotlin Toolchain 0.13 improvements for creating and configuring multiplatform projects, while the State of Kotlin in 2026 report notes Kotlin’s 15-year growth, widespread backend use, and high adoption of AI coding tools among surveyed developers. KotlinConf 2027 will take place in Kraków on April 21–23, with speaker submissions open until November 27 and a multiplatform contest offering top student and graduate entrants trips to the event. Kotlin 2.5.0-Beta2 introduces features including name-based destructuring, return-value checking, Java 28 support, expanded Swift export, and default incremental compilation for Apple targets, while upcoming releases will require JDK 17 or newer to run the compiler. The roundup additionally points developers toward Klibs.io resources for comparing multiplatform libraries and discovering tools such as Ktorfit and the Klibs.io MCP server.
Oct 30, 2026
756 words in the original blog post.
IntelliJ IDEA 2026.2.4 is available through the IDE, JetBrains Toolbox App, Ubuntu snaps, and the JetBrains website. This bug-fix update resolves issues affecting WSL project files incorrectly appearing missing after prefetch failures, progression in the Java onboarding tour, import optimization performance during formatting and commits, and IDE startup when the plugin cache contains invalid data. Users can consult the release notes for the full list of fixes and report additional problems through the issue tracker.
Oct 30, 2026
185 words in the original blog post.
JetBrains IDEs 2026.3 expands remote development by integrating remote session setup, management, and reconnection directly into the IDE, reducing reliance on JetBrains Gateway or Toolbox as separate entry points. Users can open recent remote projects from the Home screen, establish SSH connections through a two-step workflow, automatically install or update remote backends, and use remote home-directory paths with familiar tilde syntax. The release also introduces native OpenSSH support, enabling broader authentication and feature compatibility while consolidating connections from OpenSSH configurations, IDE settings, Toolbox, and recent history; OpenSSH is expected to become the default implementation in future releases. During remote backend startup, the IDE now enters Light Mode immediately, allowing users on supported Linux hosts to browse, edit, search, and use terminals before transitioning seamlessly to full Smart Mode without losing work.
Oct 30, 2026
758 words in the original blog post.
JetBrains is responding to an industry-wide eMudhra certificate authority incident reported on October 17, 2026, in which weaknesses in the generation of private keys for eight intermediate CAs could allow attackers to create apparently trusted certificates. The issue affects JetBrains products that rely on Java-based or bundled certificate trust stores, although JetBrains reports no known exploitation against its customers. Security fixes are already available for many products, including JetBrains IDEs, Gateway, Toolbox App, Rider, ReSharper, JetBrains Runtime, YouTrack, Hub, Datalore, and Junie CLI, while updates for TeamCity, Kotlin tools, IDE Services, MPS, some Qodana components, and command-line tools remain in progress. Customers are advised to install product and operating-system updates, update Java runtimes where applicable, and follow product-specific mitigation guidance; if updates cannot be applied, they may remove the affected emSign ECC Root CA – G3 and AAA Certificate Services roots from Java trust stores rather than disabling certificate validation. JetBrains is also updating its cloud infrastructure, requiring no customer action except for TeamCity Cloud users with self-hosted build agents, who should verify and update their agents’ Java runtimes.
Oct 30, 2026
1,140 words in the original blog post.
JetBrains has issued security updates after the private keys for eight eMudhra intermediate certificates were compromised, creating a risk that network attackers could impersonate websites and services accessed by JetBrains products. The issue affects IDEs, ReSharper and other .NET tools, Toolbox App, Gateway, the bundled JetBrains Runtime trust store, the JCEF embedded browser, and the separate jetbrainsd background service; operating system updates alone do not resolve it. JetBrains reports no known exploitation against its users and recommends updating affected software to specified fixed releases, particularly because versions from 2025.1 and earlier, as well as Toolbox App 3.7 and earlier, will not be patched. The fixes remove trust in compromised certificate chains, update JCEF certificate protections where available, and require jetbrainsd version 0.9.9728 or later. Users unable to update can temporarily remove affected certificates from a copied Java trust store, disable JCEF, and manually update jetbrainsd, while also updating project JDKs and independently bundled runtimes such as Node.js, Python, and Go.
Oct 30, 2026
1,315 words in the original blog post.
JetBrains Academy’s October digest highlights upcoming events, AI coding tools, and new learning opportunities, including the free online JetBrains AI.js Day 2026 on November 13, where frontend experts will discuss practical concerns about AI. It promotes JetBrains AI in IDEs and the free Junie Lite configuration, emphasizing that developers can use agents for boilerplate while reviewing generated code through diffs, tests, and debugging tools. The digest also announces its first Coursera specialization, Python Development Essentials for Science and Data, comprising introductory Python, NumPy, and pandas courses, alongside a free DeepLearning.AI course on moving AI coding workflows from cloud-based tools toward local control. In addition, it features three Udemy Full Stack Python Architect courses by Chand Sheikh, covering Django, FastAPI, or Flask with React and AI-assisted development in PyCharm.
Oct 29, 2026
497 words in the original blog post.
JetBrains Research’s Koog Agents Optimization Library is a Kotlin-based research tool for automatically improving selected prompts within Koog agent workflows without changing their control-flow graphs. Developers mark instruction-bearing subgraphs as optimizable, supply a training dataset and scoring function, and use optimizers such as GEPA to iteratively test prompt variants, analyze failures through a reflection model, and save improved instructions as a reviewable JSON artifact. In a privacy-focused PAPILLON agent evaluated on the PUPA benchmark, GEPA expanded initially vague redaction guidance into detailed privacy-preserving instructions that balanced zero personal-data leakage with useful task context, improving the average score from 81.19 to 94.22 across five runs. The library also supports BootstrapFewShot, MIPROv2, and ACE through a common training and loading interface, allowing developers to switch optimization approaches while retaining the same surrounding agent code.
Oct 29, 2026
1,319 words in the original blog post.
SEGGER’s J-Link debug probes, Ozone debugger, and SystemView real-time analysis tool have gained deeper integration with JetBrains CLion, allowing embedded developers to configure probes, program devices, debug applications, and analyze system behavior within the IDE. CLion’s latest release also adds AI-assisted hard-fault debugging for Cortex-M devices, using debugger data such as fault registers to trace hardware issues to their source and support faster diagnosis and resolution. The companies describe the integration as a step toward hardware-aware AI development, in which agents can use information from physical targets rather than relying only on source code. J-Link supports more than 15,000 microcontrollers and SoCs from over 120 vendors and works with more than 35 development environments, while a free JetBrains and SEGGER webinar is scheduled for November 10, 2026 to demonstrate the new capabilities.
Oct 29, 2026
777 words in the original blog post.
JetBrains reports that Rider, a Platinum sponsor of the Godot Foundation through 2027, has expanded into a unified IDE for Godot development across GDScript, C#, and C++ GDExtensions, while remaining free for non-commercial use. Recent releases added GDScript-aware navigation, refactoring, formatting, scene previews, node-path generation, debugging, testing support, CMake integration, native debugging visualizers, and templates for games, assets, editor plugins, and extensions. An official Rider Integration addon connects the IDE with the Godot editor, and mixed-language projects can debug GDScript and C# simultaneously. The Rider 2026.3 Early Access Program introduces code-to-scene navigation, signal-connection discovery, debugger completion, remote scene inspection for C#, and better handling of multiple Godot projects. JetBrains also describes upstream contributions to Godot’s debugging protocol, feedback-driven fixes, and emerging AI-agent integrations through its Air plugin and MCP server, while inviting developers to test preview features and provide feedback.
Oct 29, 2026
1,632 words in the original blog post.
JetBrains Air integrates coding agents into JetBrains IDEs, allowing developers to launch, monitor, and review multiple agent sessions alongside their normal coding, testing, and debugging work. Available through the JetBrains Marketplace plugin and built into the 2026.3 EAP, Air supports agents including Codex, Gemini, GitHub Copilot, Claude Agent, Junie, and other ACP-compatible tools, while leaving users in control of which providers they connect and what data is shared. Users can start sessions with keyboard shortcuts, provide context through files or selected code, track progress in editor tabs, review diffs, run tests, send feedback, and view estimated session costs. The introductory workflow emphasizes assigning focused tasks such as code exploration, bug investigation, or feature implementation, then using familiar IDE tools to evaluate whether agent-produced changes are ready to accept.
Oct 28, 2026
848 words in the original blog post.
JetBrains reports that Junie Local for Mac M5 has increased coding-generation performance from roughly 40 to 70 tokens per second, with bursts reaching 107 tokens per second, through adoption of the Splash inference engine and newly fine-tuned DFlash2 draft models. Splash improves speculative decoding by verifying multiple drafted tokens per step and reducing the context-related cost of attention, while JetBrains also ported its M5 int8 optimization from MLX-VLM to the new engine. To reduce excessive reasoning-token generation associated with Qwen 3.8, the update uses the Qwen3.8-3.6-27B-blend model and a DFlash2 drafter fine-tuned on internally generated coding-agent trajectories, achieving strong acceptance rates with about one million generated training tokens. The release also raises prefill speed from 850 to 1,200 tokens per second, reduces memory use to about 24 GB for a 100,000-token context, and expands support to M4 and newer Macs with 64 GB of RAM, although M4 prefill remains substantially slower than M5.
Oct 28, 2026
717 words in the original blog post.
JetBrains’ Junie Eco Mode reduces AI coding costs by assigning task planning to a user-selected high-capability model and implementation, testing, and iterative fixes to a lower-cost model, currently GPT-6 Luna. The selected “architect” model examines the code and produces a structured plan covering goals, scope, locations, approach, tests, and risks, while the “builder” executes that plan through a single handoff rather than repeatedly consulting a premium model. Internal benchmarks across real-world private-repository tasks found average model cost reductions of 57.3% on one benchmark and 48.1% on another, with average task-resolution rates changing by less than one percentage point, though individual configurations varied. Eco Mode applies only to code-change tasks intended to be retained and remains unavailable for remote sessions and certain model-and-effort combinations. It is enabled by default for supported configurations in recent JetBrains IDE releases, JetBrains Air EAP, and Junie CLI version 26.10.20 or later, while users can disable it when they prefer a single model to handle an entire task.
Oct 28, 2026
1,627 words in the original blog post.
Starting with version 2026.3, JetBrains IDEs replace their long-used JediTerm terminal emulator with libghostty, the high-performance library behind Ghostty, while retaining IDE-controlled rendering and integrations. The change improves compatibility with modern terminal-based tools such as AI coding agents, Neovim, tmux, fish, dashboards, and system monitors through support for protocols including Kitty keyboard handling, xterm modifyOtherKeys, theme detection and notifications, expanded color queries, and numerous previously unsupported escape sequences. Users should see reliable multiline prompts and modifier keybindings, automatic dark and light theme synchronization in compatible applications, and better rendering of borders, colors, repeated characters, and other rich terminal output. Libghostty also provides groundwork for future capabilities such as remote clipboard access, task-completion notifications, progress bars, images, and styled underlines. The new engine is enabled by default in 2026.3 EAP builds and can be disabled through an IDE registry setting if problems occur.
Oct 27, 2026
1,029 words in the original blog post.
YouTrack presents AI as an optional component of its project-management platform, allowing teams to use external AI tools through its Model Context Protocol (MCP), enable a built-in AI Assistant, or disable AI entirely while retaining full core functionality. MCP integrations support AI chats, coding agents, IDEs, automation platforms, and internal workplace applications, with permission-aware access to YouTrack data through OAuth, automatic client registration where supported, or permanent tokens for compatible use cases. The built-in assistant can create issues from selected text, improve writing, summarize content, suggest replies, and extract text from images, while JetBrains states that customer data is not used to train AI models. Planned Q4 2026 additions include Ask YouTrack, an agentic chat for searching projects and proposing task updates, bring-your-own-key support for customer-selected language models, and AI-assisted task creation from Slack threads. Available in self-hosted Server and JetBrains-managed Cloud versions, YouTrack also emphasizes infrastructure and data-control choices, compatibility with teams that prohibit AI, and pricing positioned as potentially lower than combined Jira and Confluence costs.
Oct 26, 2026
1,988 words in the original blog post.
JetBrains’ Q3 2026 newsletter for IntelliJ Platform plugin developers highlights the free online JetBrains Plugin Developer Conf on November 10, featuring sessions on AI integrations, platform updates, scalable plugin architectures, and practical technologies such as LSP, WSL, Dev Containers, and web UIs. Marketplace improvements now allow plugin descriptions to be edited independently of plugin uploads and provide UI freeze reports with thread dumps and CPU snapshots to help developers diagnose performance issues. Tooling releases include IntelliJ Plugin Verifier 1.411, which improves content-module support, reporting, APIs, and memory efficiency, and IntelliJ Platform Gradle Plugin 2.19.0, which adds multi-platform distribution builds, offline support, testing options, updated verification behavior, and several fixes alongside removal of a deprecated task. The newsletter also recommends release channels for safely testing prerelease versions with limited users and points readers to resources on the new web-based IDE Plugin Generator, faster native WSL project support, the evolving role of IDEs in agentic development, migration from JNA to Java’s Foreign Function and Memory API, and simple persistent plugin settings with PropertiesComponent.
Oct 26, 2026
798 words in the original blog post.
JetBrains has published YouTube recordings from GameDev Day 2026, held on September 28, featuring technical talks, demos, and discussions for game developers working with Unreal Engine, Unity, Godot, game engines, infrastructure, and cross-language systems. Sessions address subjects including community feedback and AI, TeamCity-based CI workflows in Unity, gradual adoption of ECS in live games, static analysis for AI-generated code, AI-assisted procedural generation in Godot, Unreal performance profiling, Figma-to-Unity UI automation, agentic development in Rider, Epic’s Lore VCS for large binary assets, and debugging C# clients alongside C++ servers. The event included online attendees and watch parties in Belgrade and Limassol, with JetBrains encouraging viewers to explore its game-development tools, including Rider, which is free for non-commercial use, and to subscribe for future GameDev Day updates.
Oct 23, 2026
1,083 words in the original blog post.
JetBrains has made Air Governance, formerly JetBrains Central, generally available as the organizational management layer for its Air agentic-development product suite. The platform centralizes oversight of AI agents, models, providers, access policies, and spending across tools including Claude Agent, Codex, Gemini CLI, Junie, and other ACP-compatible agents used through IDEs, command-line tools, browsers, and mobile devices. Administrators can set credit limits, approve agents and MCP servers, use JetBrains AI or bring their own provider accounts, and apply policies for teams or users. Air Governance also provides cost and outcome analytics, including spending by user, agent, model, and provider, alongside metrics on AI-generated and AI-modified code. JetBrains developed the service after its own AI spending rose about tenfold in early 2026 and now uses it to oversee roughly 2,300 monthly users. Available under the Business plan for $20 per AI-enabled developer per month, the platform will gain expanded MCP propagation, GitHub Copilot and cloud-agent connections, advanced analytics, BYOK pricing controls, and team skill management in future releases.
Oct 22, 2026
789 words in the original blog post.
As coding agents increasingly handle multi-file tasks, dependency exploration, and refactoring, the author argues that IDEs remain essential as integrated environments for understanding project context, directing agent work, verifying changes, and maintaining accountability for shipped software. The piece compares the IDE to an aircraft cockpit: agents can execute substantial work, but engineers must still judge architectural fit, review consequences, diagnose failures, and take responsibility for decisions. It contends that IDE capabilities such as code navigation, type checking, linting, testing, and dependency analysis should support both developers and agents, grounding probabilistic AI output in deterministic validation. The author also advocates avoiding dependence on a single AI provider through interoperable standards such as the Agent Client Protocol and Bring Your Own Keys, allowing developers to select and change models without rebuilding workflows. Overall, it presents agentic development not as a replacement for engineering judgment, but as a shift from manually producing code toward defining intent, evaluating outcomes, and solving broader software problems.
Oct 20, 2026
3,165 words in the original blog post.
JetBrains Plugin Developer Conf 2026 is a free online event scheduled for November 10 at 1:00 pm UTC, focused on current tools, APIs, and practices for developing plugins for JetBrains IDEs. The program will examine AI-assisted development and its limitations, native LSP support for language plugins, web-based profiler integration, scalable integration architecture, and EelApi support for varied project environments. Speakers from JetBrains and the broader developer community will present technical talks, practical examples, and live Q&A sessions, including an opening keynote and sessions on plugins such as Dependency Assistant. Attendees may join the full event or individual presentations, with live YouTube streaming and recordings available afterward.
Oct 20, 2026
303 words in the original blog post.
Ulises Olave, CTO and co-founder of Peru-based Fitia, describes how the AI-powered meal-planning and calorie-tracking app has grown to more than 10 million users while processing over 30 million food logs each month alongside health data such as weight, sleep, steps, and water intake. As its data became distributed across PostgreSQL, BigQuery, Azure, and Redis, Fitia adopted JetBrains DataGrip to centralize database work and use AI-assisted query suggestions, which the team says reduced some tasks from hours to seconds. The company also introduced GoLand for its Go microservices, reporting that its small engineering team roughly doubled development speed and now delivers a major feature about once a month, three times faster than previously. Looking ahead, Fitia plans to build personalized nutrition features using large-model providers’ capabilities combined with its own community data as it expands in the United States. The accompanying JetBrains Startup Program information states that privately owned companies established within the past five years can apply, with reviews generally completed within 48 hours and eligible participants receiving access to selected IDE, team, and learning tools.
Oct 20, 2026
1,112 words in the original blog post.
JetBrains Research argues that coding-agent benchmarks should reflect the task flows of specific user populations rather than treat issue-based bug-fix and feature tasks as universally representative. Analysis of 4,782 IDE agent sessions found that developers frequently ask for explanations, planning, reviews, refactoring, debugging, and command execution alongside change requests, whereas SWE-Bench Pro consists almost entirely of single-message feature and bug-fix tasks. To address this gap, the researchers introduce SWE-TaskFlow, a method that preserves existing benchmark repositories, tests, and requirements while splitting issues into sequential requests and inserting verifiable repository questions, then selecting message sequences that align with observed session lengths, intents, and transitions. Applied to 700 SWE-Bench Pro tasks, the approach improved alignment with long production sessions and enabled intermediate evaluation checkpoints. Pilot experiments across four models showed that sequentially revealing requirements substantially increased cost while producing only modest changes in resolve rates, and added questions imposed further overhead while measuring repository-question accuracy. The authors note that the method does not yet model many real interactions or reactive user feedback, but position it as a reusable framework for calibrating benchmarks to declared target workflows.
Oct 20, 2026
2,626 words in the original blog post.
JetBrains .NET Day Online 2026, held on October 7, has made recordings of all sessions available on YouTube, covering AI-assisted development, distributed systems, performance, security, architecture, interoperability, databases, and cloud deployment. The keynote introduced JetBrains Air for Rider, showing how agents can use IDE capabilities for tasks such as testing, debugging, refactoring, code review, and profiling, while also noting ReSharper options for VS Code and Cursor and the availability of Junie Lite. Other presentations examined messaging patterns including Outbox, Inbox, and Saga; C# interceptors and source generators for Native AOT; AI development workflows; .NET and NuGet supply-chain risks; modular monolith architecture; cross-platform .NET development and MCP servers; Microsoft’s SQL MCP Server; and scaling Blazor applications with .NET Aspire and Azure Container Apps. Attendees are invited to provide feedback through a survey open until November 8, with support for the event provided by the .NET Foundation.
Oct 19, 2026
824 words in the original blog post.
JetBrains Research reports a rapid rise in developers’ willingness to delegate software-development tasks to AI agents, with readiness to hand over application-logic coding increasing from 18% in April–June 2025 to 50% a year later, test writing from 50% to 75%, and refactoring from 38% to 60%. Developers remain more cautious about fully delegating code review and debugging, although acceptance is growing in both areas. Sentiment toward AI tools is also shifting, as 79% of respondents expect employers to require AI-development-tool proficiency, while concerns about reduced demand for professional developers and eventual full replacement of human coders have increased, particularly among frequent users of coding agents. The findings draw on JetBrains’ localized and statistically weighted Developer Ecosystem and AI Pulse surveys, covering coding-related roles worldwide.
Oct 19, 2026
649 words in the original blog post.
Researchers compiled the decision-making logic of a limited Maven-compatible build controller from restricted Python into fixed neural-network weights, rather than training a model, and used the resulting network to build ZXing 3.5.3 with byte-identical JAR outputs to Apache Maven. The network executes its encoded instructions sequentially through matrix operations and relies on a non-decision-making bridge for file access and external tools such as javac, JUnit, bnd, and zip; it passed hundreds of handwritten cases, 10,000 generated project tests, 33 additional projects, and ZXing’s 549 upstream tests. Although a full clean build with tests was initially 1.49 times slower than Python, compiler improvements reduced the weight table from 29,625 to 26,626 rows and execution steps by 23%, bringing the full-build overhead to about 1.42 times, while GPU execution remained slower because the workload offers little parallelism. A separate maze-solver experiment indicated the approach generalizes beyond build logic but is substantially slower than ordinary code. The authors position the work not as a Maven replacement or machine learning system, but as a reproducible, interpretable network with known weight meanings that could support interpretability research, exact neural subroutines, and hashable, replayable decision policies, while noting that much of Maven functionality and several plugins were outside the project’s scope.
Oct 18, 2026
2,403 words in the original blog post.
JetBrains has expanded Junie Local, its free local AI coding agent, with day-one support for NVIDIA RTX Spark systems and stable Windows support for NVIDIA RTX GPUs with at least 24 GB of VRAM. Running entirely on-device, Junie Local can read and edit code, execute tests, and avoid token charges, quotas, and external code transmission. RTX Spark’s Blackwell GPU and Grace CPU enable local model inference speeds of up to 55 tokens per second in JetBrains’ initial testing, while adaptive context sizing, FP8 Tensor Core computation, and speculative decoding improve performance. The updated installer automatically checks compatible hardware and drivers, installs an optimized engine and model, and supports Qwen-based model options; users need Windows 11, NVIDIA driver version 581 or later, and roughly 20 GB of disk space. JetBrains plans further model, memory-efficiency, coding-performance, and RTX Spark inference improvements, and is inviting feedback from users on supported hardware.
Oct 16, 2026
765 words in the original blog post.
JetBrains has released version 1.0 of its Java and Kotlin by IntelliJ IDEA extension for VS Code, Cursor, and other LSP-compatible editors, following a preview period and collaboration with Cursor. Powered by IntelliJ IDEA’s language technology, the extension provides Java, Kotlin, and mixed-language project support, code completion, navigation, analysis, refactoring, Maven, Gradle, and Bazel integration, JUnit support, and debugging features designed to work reliably on large projects. The release reflects JetBrains’ effort to make JVM development intelligence available beyond its desktop IDE and to support agentic development, where language servers can provide faster and more deterministic code analysis with lower token use. Cursor includes dedicated Agents Window support and IntelliJ-based inspections, including nullability analysis and Spring-specific capabilities. The extension is available through the Visual Studio Marketplace and Open VSX, but now requires an IntelliJ IDEA Ultimate subscription after a 30-day free trial; JetBrains recommends disabling overlapping Java extensions from Red Hat and Oracle. Future updates are expected to let agents directly call the language server in Cursor, VS Code, and terminal-based tools such as Claude Code or Codex.
Oct 15, 2026
984 words in the original blog post.
Google Summer of Code 2026 marked the Kotlin Foundation’s fourth year in the program, with two completed projects focused on Kotlin/Wasm performance and Kotlin education worldwide. Ayako Hayasaka added WebAssembly tail-call support to the Kotlin/Wasm backend, enabling recursive programs to avoid call-stack growth, improving benchmark performance by 1.7 to 4.3 times at a depth of 10,000, and allowing programs to complete at one million recursive calls where they previously crashed; her investigation also identified and fixed a V8 engine optimization issue. Mark Seif Khalil developed an automated, structured approach for collecting and organizing fragmented information about where and how Kotlin is taught, creating a database and reusable foundation for monitoring educational adoption, trends, and gaps. The Kotlin Foundation thanked the contributors and mentors and encouraged prospective future contributors and mentors to join its GSoC community channel.
Oct 15, 2026
654 words in the original blog post.
Netflix integrated JetBrains’ headless Java/Kotlin code-intelligence engine, based on IntelliJ IDEA’s indexing capabilities, with Claude Code and Codex to help coding agents navigate JVM projects and compiled JAR dependencies more accurately. Because the LSP server alone was not suited to repeated agent calls, Netflix created a Java-based daemon and CLI that keep project indexes warm, avoiding repeated startup and indexing costs. Across benchmarks on an internal Spring Boot service and the open-source DGS Framework, agents using the intelligence layer generally consumed less time and fewer tokens while maintaining or improving correctness, with typical efficiency gains of 20–30% and a best case near 65%, especially for stronger models operating at higher effort settings. End-to-end programming tasks showed consistent gains of roughly 10–30%, while navigation tasks reduced token use by about 15–60%, although one batched-search configuration did not show representative time savings. Netflix rejected alternatives requiring a running IDE or relying on Eclipse’s Java LSP, plans to test similar methods for other languages, and JetBrains argues that future agent tooling should expose program-level concepts such as symbols, diagnostics, tests, and runtime behavior rather than primarily relying on text-position-based LSP interactions.
Oct 15, 2026
2,043 words in the original blog post.
KotlinConf 2027’s call for proposals encourages developers of all experience levels to submit talks rooted in useful Kotlin-related lessons, including successes, failures, migrations, technical discoveries, and practical solutions. Strong submissions present a focused story rather than a broad subject, using a clear, informative title and a concise abstract that explains the context, challenge, approach, and concrete audience outcomes. Reviewers look for evidence of care, clarity, authenticity, and a direct connection to Kotlin, while specific takeaways, a distinct perspective on familiar topics, and honest discussion of trade-offs can make proposals more compelling. Speakers should identify the intended audience and appropriate format, choosing lightning talks for narrower ideas and regular sessions for material requiring more depth, while recordings or other relevant work can help demonstrate communication style. Before submitting, applicants are advised to seek feedback, verify that the proposal is realistic and polished, and ensure that AI tools have not replaced their own voice or experience.
Oct 15, 2026
3,511 words in the original blog post.
JetBrains has released the MPS 2026.2 Release Candidate for testing ahead of the official launch, adding numerous bug fixes and improvements over the earlier EAP1 build. The release introduces Linux ARM64 support for MPS and standalone MPS-based applications, extends Build Language with TAR archive imports, and preserves symbolic links and permissions in platform-specific packaging, including a move to .tar.gz archives for macOS distributions. The Projectional Agent Toolkit receives substantial MCP API improvements intended to reduce AI-agent token usage, alongside new Typesystem and other MCP tools for type analysis, project management, logs, intentions, quick-fixes, and Java expression correction. Additional updates rework the Import dialog to respect user selections and improve its controls, while refinements to Diff dialogs and editor markers enhance alignment, resizing behavior, visual consistency, and hint display.
Oct 15, 2026
892 words in the original blog post.
Verifone architect Yuri Geronimus’s KotlinConf 2026 presentation argues that mission-critical payment systems become safer when code mirrors regulatory and card-network specifications closely enough for the Kotlin compiler to enforce their rules. Operating across terminals, gateways, banks, and regulators requires handling thousands of scenarios, offline operation, conflicting national requirements, sensitive card data, and transaction states where small semantic errors can cause outages, compliance breaches, fraud, or duplicate charges. Geronimus advocates patterns such as value classes for distinct identifiers, boundary validation with sealed results, non-null mandatory fields, masked sensitive-data wrappers, immutable state-specific models, exhaustive branching without fallback cases, structured concurrency chosen according to required failure behavior, and timeouts treated as explicit business outcomes. He also recommends readable Kotlin DSL-based certification tests so non-engineering specialists can review them against specifications. Verifone reported faster development, fewer certification defects, a smaller codebase, and quicker onboarding after adopting these approaches. Rather than rewriting legacy Java and C++ payment platforms, Geronimus recommends parallel modernization through structural improvements, acceptance testing, compliance and reliability fitness functions, and incremental Kotlin adoption, framing migration as a business investment requiring an operational plan rather than simply a language preference.
Oct 14, 2026
2,847 words in the original blog post.
Lauterbach and JetBrains announced an integration between TRACE32 debugging tools and the CLion C/C++ IDE that lets embedded developers debug applications running on target hardware without leaving their development environment. Using the Debug Adapter Protocol and the TRACE32 Debug Adapter, developers can launch a new TRACE32 PowerView session or attach CLion to an existing instance, then control execution, step through code, and inspect variables while TRACE32 manages hardware communication. The feature requires CLion 2026.2.4 or later, TRACE32 PowerView 09.2026 or build 195402, and TRACE32 Debug Adapter 0.1.2 or later. The companies position the integration as an extension of CLion’s coding, navigation, testing, and embedded workflows alongside TRACE32’s debugging, tracing, performance analysis, and code-coverage capabilities across the embedded product life cycle. JetBrains and Lauterbach will demonstrate the integration in a free webinar on October 27, 2026.
Oct 14, 2026
448 words in the original blog post.
JetBrains Air’s cloud agents allow developers to run AI coding tasks remotely in managed, isolated environments rather than on local machines, with organizational availability currently offered through Air IDE integrations and Air Teams. The discussion emphasizes sandboxing as a means of limiting an agent’s access to files, networks, APIs, secrets, and system resources, reducing risks such as unintended changes, data exposure, and malicious prompt-driven behavior. It compares local workspaces and Git worktrees, which can share aspects of a developer’s machine, with remote SSH, Docker, Docker Sandboxes, and cloud execution, which provide progressively stronger isolation but differ in required setup and administration. In Air, each task receives a new cloud VM containing a Docker container, with all agent network traffic routed through a proxy that can restrict connections to trusted source-control and package-registry domains, permit selected additional domains, allow all internet access, or disable it entirely. Secrets are encrypted and supplied as environment variables without exposing shared values to teammates, while repository access remains limited to the initiating developer’s existing permissions.
Oct 14, 2026
1,095 words in the original blog post.
Qodana extends IntelliJ IDEA’s static code inspection engine into CI/CD pipelines, allowing teams to identify code-quality issues early, enforce consistent standards, and use quality gates to prevent problematic changes from reaching production. The article reviews six supported platforms: GitHub Actions, which offers straightforward marketplace integration and pull-request support; GitLab, whose native CI/CD component integrates reports into merge requests; TeamCity, which provides a native GUI-based runner and built-in reporting; Jenkins, which uses Docker-based pipeline configuration for highly customizable enterprise workflows; Azure Pipelines, which uses the QodanaScan task and displays results in an Extensions tab; and CircleCI, which provides a container-optimized orb. Each option has trade-offs involving configuration complexity, costs, platform requirements, scalability, or flexibility, but all aim to make automated static analysis a routine part of commits and code reviews without requiring teams to change their preferred CI/CD environment.
Oct 14, 2026
1,390 words in the original blog post.
CI/CD pipeline observability enables teams to understand not only what failed or slowed down in delivery pipelines, but why, by connecting build results, metrics, logs, events, and traces across repositories, agents, tests, infrastructure, and external services. Key signals include queue and stage durations, success and retry rates, flaky tests, cache performance, agent utilization, resource consumption, costs, and change metadata, with detailed breakdowns and percentile measurements often revealing issues hidden by averages or aggregate success rates. Built-in CI/CD dashboards, used by 76% of surveyed developers, provide a practical starting point because they retain native pipeline context, while Prometheus and Grafana support infrastructure metrics, Elastic Stack and OpenSearch centralize searchable logs, OpenTelemetry standardizes portable collection of telemetry for external backends, commercial platforms provide managed cross-system analysis, and custom solutions address specialized requirements. Selecting an approach depends on investigation depth, preservation of CI/CD context, correlation across systems, setup and maintenance effort, developer usability, security and data residency needs, cost, and portability. Most organizations benefit from combining tools rather than choosing a single platform, beginning with existing CI dashboards and adding metrics, logging, or tracing capabilities only where they address real troubleshooting gaps.
Oct 13, 2026
3,353 words in the original blog post.
At AGNTCon and MCPCon Europe, discussions among open-source and agentic AI participants focused on governance, trust, funding, and the growing need for human accountability in AI-assisted development. Speakers emphasized that while financial support sustains project infrastructure, decision-making legitimacy depends on consistently earned trust and transparent contribution. AI-generated proposals and code are increasing the burden on maintainers and reviewers, as polished but generic submissions, repeated debates, faulty system connections, undetected job failures, and fabricated data can obscure serious problems. Presentations argued that code review, regression testing, planning, and human oversight remain essential because agent-produced output can appear complete while being incorrect. The event also highlighted opportunities for broader participation in emerging standards for agent-tool connections, tracing, security, governance, and accountability, noting that these systems must support diverse models and vendors rather than a single proprietary stack.
Oct 13, 2026
842 words in the original blog post.
JetBrains Academy outlines its 2026–2027 activities for school students interested in mathematics, programming, and artificial intelligence, including free online Youth Math, AI, and Coding Clubs, with Math and AI registration open until October 31 and Coding Club sessions beginning October 22. In-person JetBrains Coding Cups for students aged 13–19 will take place in Cyprus on October 17, Poland on October 24, and Serbia on November 14, while the fourth Cyprus Programming Challenge is scheduled for November 8. The Youth Coding Challenge on November 29 will include junior and senior leagues, potentially leading junior winners to advanced Coding Club participation and allowing senior results to contribute to selection for the ACTS 2027.1 camp. Held in Cyprus from March 15 to 25, 2027, ACTS will offer Competitive Programming, Software Engineering, and AI tracks, combining intensive instruction, practical challenges, contests, and peer collaboration; its AI track will be invitation-only for national teams from five countries and selected Cyprus-based students.
Oct 13, 2026
828 words in the original blog post.
Python and Go serve overlapping but distinct development needs: Python emphasizes readable, flexible syntax, rapid iteration, and a vast ecosystem for machine learning, data science, scientific computing, and prototyping, while Go prioritizes static typing, explicit error handling, standardized formatting, native compilation, concurrency, and maintainability in large services. Drawing on JetBrains’ 2026 survey of nearly 29,000 respondents and interviews with seven recent Python-to-Go adopters, the comparison finds that Go developers tend to be more experienced and infrastructure-focused, commonly building APIs, cloud services, and DevOps tooling around platforms such as Kubernetes, whereas Python has a larger beginner and early-career population and remains dominant in AI model development and analytics. Go’s performance, fast startup times, straightforward deployment, goroutines, and compile-time checks motivate migrations for latency-sensitive or high-concurrency workloads, although developers often find its verbosity, repetitive error handling, and less flexible testing model challenging. Python’s libraries mitigate interpreter-speed limitations for numerical and ML workloads, while emerging free-threaded builds may reduce constraints imposed by the Global Interpreter Lock. Rather than replacing Python entirely, teams commonly combine the languages, using Go for production services and infrastructure while retaining Python for data pipelines, notebooks, model training, and scientific libraries.
Oct 13, 2026
8,710 words in the original blog post.
JetBrains Academy is launching an October Build Your Game Challenge that invites participants to complete its Build a Game With an AI Coding Agent course, enhance the resulting game with original features, and submit it through the community Discord server. The challenge begins October 12, submissions and community voting open October 19, both close October 23, and winners will be announced October 28. Entrants may use any development setup, with JetBrains promoting its Air IDE integration and free Junie Lite coding agent as an accessible option, while allowing stronger agents or personal AI subscriptions for more ambitious work. Winners will be selected through community reactions and a JetBrains jury evaluating creativity, originality, gameplay, and player experience, with digital gift-card prizes ranging from four $50 jury awards to a $400 Community Choice first prize.
Oct 12, 2026
624 words in the original blog post.
JetBrains Marketplace now provides IntelliJ Platform plugin authors with UI freeze reports attributed to their plugins through a Freezes tab, provided they register the built-in Marketplace error handler in plugin.xml. Reports are grouped by similar error type and include a representative thread-dump excerpt, recurring full dumps captured during freezes, and sometimes CPU snapshots, stack analyses, and OpenTelemetry performance metrics. Investigation should begin with the event dispatch thread to determine whether it is performing expensive work or blocked by another thread, while comparisons across multiple dumps can reveal persistent operations or lock contention, including long non-cancellable background read actions. JetBrains also offers an analyze-freeze skill for AI coding agents that can help trace threads, coroutines, services, and likely plugin causes, although developers should follow diagnostic-data policies and verify all conclusions against full dumps and source code. Authors are advised to confirm the affected plugin version, avoid assumptions when version data is missing, use cancellable and short read actions where appropriate, keep blocking I/O outside read actions, validate fixes with fresh diagnostics, and seek help in the JetBrains Platform Forum when needed.
Oct 09, 2026
1,089 words in the original blog post.
JetBrains’ State of CI/CD in 2026 report combines findings from four studies to examine pipeline usage, infrastructure choices, tool adoption, and AI’s evolving role in software delivery. Among DevOps practitioners surveyed, 83.4% use CI/CD pipelines, while 72.8% operate their primary systems on premises or through self-managed cloud infrastructure; 30.5% specifically use on-premises installations, the highest proportion in five years. AI is used for at least one CI/CD-related task by 48.8% of organizations, though only 8% embed AI-powered steps directly in builds or tests, with more common applications including code review, test creation and repair, and build-failure diagnosis. The report also addresses leading CI/CD tools, the reasons organizations use multiple platforms or consider switching, and barriers limiting broader AI adoption.
Oct 08, 2026
264 words in the original blog post.
Adapted from a JetBrains Game Development Day 2026 talk, the post argues that code quality in game development encompasses correctness, performance, stability, security, maintainability, and reusability, all of which can affect player experience, reviews, revenue, and long-term project costs. It notes that delivery pressure and widespread AI-assisted coding can accelerate prototyping and reduce repetitive work, but may also rapidly scale problems such as logic errors, duplicated code, outdated dependencies, and security vulnerabilities. The proposed approach combines deterministic static analysis with AI-based review: static tools identify rule-based issues, enforce standards, scan dependencies, and provide reliable automated fixes, while AI can help assess intent, cross-project logic, refactoring, and more complex corrections. Using Qodana as an example, the post describes integrating analysis into CI pipelines, baselining existing issues, focusing pull-request scans on changed files, and feeding AI-generated changes back through static checks before merging. It also cites studies suggesting that combining static analysis with LLMs can reduce token use and vulnerabilities, while cautioning that emerging AI-related data should be critically evaluated.
Oct 08, 2026
2,511 words in the original blog post.
JetBrains has released Mellum2.1, an Apache 2.0-licensed 12B mixture-of-experts open model with 2.5B active parameters designed for coding agents and self-hosted AI tools. While its architecture remains unchanged from Mellum2, the new version received extensive post-training through reinforcement learning across millions of sandboxed runs and thousands of real environments, improving its ability to explore repositories, modify files, and validate changes. Training incorporated filtered tasks spanning software engineering, tool use, math, science, and competitive programming, with an emphasis on removing unreliable or unsuitable data. JetBrains reports gains over Mellum2 and comparable open models in agentic coding, general coding, reasoning, tool calling, and general knowledge, while retaining its prior speed advantages; multi-token prediction is said to make single requests about 1.6 times faster and improve throughput under load. Mellum2.1 is available through Hugging Face, with additional builds and speculative-decoding support planned for llama.cpp, Ollama, LM Studio, and vLLM.
Oct 08, 2026
601 words in the original blog post.
IntelliJ IDEA’s command completion feature provides a context-aware way to access IDE actions directly from the editor by typing two dots, reducing the need to remember shortcuts, search menus, or switch tools. In IntelliJ IDEA 2026.3 EAP, it adds faster access to debugger logpoints, including suggestions to replace temporary print statements, and enables in-context AI actions such as explaining code, generating documentation, and refactoring selected elements with relevant file and caret information automatically supplied. The update also unifies the interface for command completion, postfix completion, and live templates through improved filtering and previews, while extending support to binary files, decompiled code, properties files, Spring workflows, and other structured environments. Intended as a discoverable complement to keyboard shortcuts, the feature analyzes code context, scope, inspections, and available actions to suggest relevant next steps in Java, Kotlin, and Scala files.
Oct 07, 2026
771 words in the original blog post.
Kotlin Toolchain 0.13 is presented as the recommended starting point for most new Kotlin Multiplatform projects, providing a unified CLI and declarative configuration for building JVM, Android, iOS, web, desktop, and server-side applications. Following its Alpha promotion and community feedback, the release improves project creation with the interactive `kotlin new` command, streamlines iOS setup through automatic simulator and runtime provisioning, adds download progress reporting, and automatically installs missing Android SDK components during IDE import. It also introduces IDE sync-status visibility, Swift Package Manager dependency support for Apple targets, improved multiplatform library publication metadata, and Kotlin/Native compiler caches intended to accelerate incremental builds. The toolchain emphasizes support for AI-assisted development through agent skills and a Hot Reload MCP server, while future plans include Homebrew installation, built-in AI skills, Swift Export support, and more automated app-store publishing workflows.
Oct 07, 2026
1,171 words in the original blog post.
TeamCity’s Warm Agents plugin reduces build queue delays by maintaining a configurable number of idle, pre-started cloud build agents for specific cloud images, provisioning replacements whenever the idle count falls below the target. Available for TeamCity 2024.12.3 or later and compatible with any supported cloud provider, it is installed through the JetBrains Marketplace and configured through REST API calls or the teamcity-cli. Organizations must balance faster build starts against the cost of cloud instances and build-agent licenses, while TeamCity respects configured limits on total agents. Targets can be adjusted dynamically, including through scheduled builds and Kotlin DSL to scale capacity up during working hours and down overnight, though reducing a target does not terminate already running agents. The plugin also exposes Prometheus-format utilization and saturation metrics to help teams analyze demand patterns and refine their agent capacity settings.
Oct 07, 2026
887 words in the original blog post.
Engineering organizations are rapidly adopting AI agents, with many developers using them daily, but faster individual coding has not consistently translated into faster team delivery. Interviews across industries identify three recurring obstacles: agent practices and configurations are often isolated, inconsistently shared, and unevenly understood; increased implementation speed shifts bottlenecks to planning, code review, and verification; and automating routine tasks introduces infrastructure, maintenance, ownership, and governance challenges. Teams may produce more code and pull requests, yet reviewers and product processes can struggle to evaluate and prepare work at the same pace, potentially reducing rather than improving overall efficiency. The author argues that effective team-wide AI use requires shared context, reusable skills, review processes suited to agent-generated output, and collectively managed automation infrastructure, themes that will be explored in future posts promoting JetBrains Air Teams.
Oct 07, 2026
1,277 words in the original blog post.
JetBrains has joined the Linux Foundation’s Open Source Security Foundation (OpenSSF) as a general member, announced alongside other new members at OpenSSF Community Day Europe in Prague. OpenSSF supports software supply-chain security through projects such as Sigstore and SLSA, while its AI/ML Security Working Group develops guidance for securing machine-learning models and AI coding assistants. JetBrains says the growth of coding agents makes software verification and provenance increasingly important, since AI-generated code and dependencies introduce new trust and security concerns. The company plans to contribute its experience from JetBrains Air, its work on helping developers understand and verify agent-produced code, while collaborating with the broader OpenSSF community to help shape secure software development practices.
Oct 06, 2026
417 words in the original blog post.
Research by JetBrains’ Human-AI eXperience team and Lund University argues that reviewing AI-generated code is primarily a trust-calibration challenge rather than a conventional diff-reading task, because language models can produce large, multi-file changes while offering no reliable indication of which portions are uncertain or risky. Based on participatory workshops with 17 practitioners and a survey of 43 software professionals, the proposed framework recommends a three-level workflow in which reviewers first gain a high-level overview, then assess risk at the file level, and finally inspect selected code snippets in detail. This approach aims to help developers allocate attention according to segment-level risk instead of auditing every generated line, while restoring contextual signals that are normally available when reviewing a human colleague’s work. Existing tools such as CodeRabbit, Claude Code, Graphite, and GitHub’s AI review features offer partial support through summaries, severity tags, or smaller review units, but the researchers identify a gap in tools that explicitly guide reviewers from overview to risk stratification to detailed analysis.
Oct 06, 2026
1,329 words in the original blog post.
Amper has been renamed and evolved into the Kotlin Toolchain with the Alpha release of version 0.11, providing a unified entry point for Kotlin development. Future Kotlin Toolchain news and release announcements will be published on the Kotlin Blog to consolidate ecosystem updates, while the existing Amper blog will remain available as an archive with its posts and links preserved but will no longer receive new content. Readers can follow future developments by subscribing to relevant Kotlin Blog topics and monitoring the project’s GitHub repository for releases and changelogs.
Oct 06, 2026
163 words in the original blog post.
JetBrains has released a Release Candidate for Java and Kotlin by IntelliJ IDEA, an LSP-based extension that brings IntelliJ IDEA’s project-aware Java and Kotlin capabilities to VS Code, Cursor, and other compatible editors. Following a public preview and several intermediate builds, the extension has surpassed 20,000 downloads and incorporates feedback from early adopters as well as internal testing on large IntelliJ-based repositories. It offers smart completion, navigation, inspections, quick-fixes, code analysis, project-wide refactorings, DAP-based debugging, Maven, Gradle, and Bazel support, test execution, and Spring tooling interoperability. In Cursor, its language server can also support coding agents with semantic search, navigation, and analysis. The RC is the last Early Access Program build and is free to use for 30 days; after the stable 1.0 release, users will receive a 30-day trial before an IntelliJ IDEA Ultimate subscription is required.
Oct 06, 2026
578 words in the original blog post.
JetBrains’ guide for content creators emphasizes that expanding opportunities in digital and tech content, including AI-enabled tools and growing audience demand, make community building increasingly important alongside content production. It encourages creators to clearly showcase their expertise, partnerships, promotional offers, and community links in published content; experiment with relevant tools through the JetBrains All Products Pack; and use active social media distribution to reach audiences beyond primary platforms. The guide also presents direct audience communication, challenges, giveaways, and milestone activities as central ways to strengthen engagement, noting that program participants may receive up to 12 free IDE subscriptions for giveaways. New creators are invited to review the Content Creators Program’s eligibility requirements and available benefits.
Oct 06, 2026
693 words in the original blog post.
JetBrains’ October 2026 dotInsights newsletter, curated by Rachel Appel, highlights .NET and broader software development resources while noting that System.Threading.Interlocked enables lock-free atomic updates to shared variables. Its collection of community links covers evolving C# features, .NET MAUI interfaces, ASP.NET Core authentication and middleware, serialization, performance, object pooling, testing, real-time systems, design patterns, extension types, Microsoft Graph migration, and AI-related topics such as RAG, structured output, agent skills, database design, and developer productivity. The issue also includes lighter social-media posts, promotes the .NET Day Online 2026 event on October 7, and shares JetBrains product news including Rider 2026.2.3 and Rider and ReSharper 2026.2.2 releases, along with an account of startup-performance improvements.
Oct 05, 2026
577 words in the original blog post.
Beginning with IntelliJ IDEA 2026.3 and Android Studio Rabbit 2 | 2026.2.2, JetBrains will discontinue Swift language IDE support in the Kotlin Multiplatform plugin, removing Swift editing, syntax highlighting, and navigation between Swift and Kotlin code. iOS application execution, Kotlin debugging on iOS, Kotlin/Native capabilities such as Swift Export, and Swift package imports will remain supported. The company cited declining adoption, developer preference for Xcode when working with Swift, and the substantial maintenance required to keep inherited AppCode-based tooling compatible with evolving Swift, Xcode, and IntelliJ Platform releases. Existing Swift functionality will remain in older, unsupported versions, while JetBrains plans to focus future investment on Kotlin tooling and the broader Kotlin Multiplatform iOS development experience rather than replacing the retired features.
Oct 05, 2026
364 words in the original blog post.
JetBrains has released TeamCity On-Premises 2026.1.5 and 2026.2.1, two maintenance updates that collectively address more than 40 security vulnerabilities alongside several performance and functional issues. Fixes include restoring TFS project diff displays, enabling Pipelines to import YAML configuration from a main repository branch, and detecting MSBuildTools after the September 2026 Visual Studio Build Tools update. Users are strongly encouraged to upgrade promptly for improved security, stability, integration compatibility, and build performance. Bug-fix releases within the same major TeamCity version use a shared data format, allowing upgrades or downgrades without backup restoration, and updates are available through automatic updates, direct downloads, or Docker images.
Oct 05, 2026
269 words in the original blog post.
October 2026’s Java Annotated Monthly highlights September’s major developments, led by the release of Java 27 and JetBrains’ announcement of JetBrains Air, an agentic software development product system now available as an IDE early-access preview. The issue features commentary from Java Champion and JetBrains advocate Marit van Dijk, along with resources on Java 27 support in IntelliJ IDEA, performance and security updates, upcoming JDK 28 proposals, modern Java development, memory management, Maven, structured concurrency, debugging with logpoints, and agent-oriented development using Java, Spring Boot, and Spring AI. It also covers Kotlin news, including its developer survey and the upcoming KotlinConf 2027, while its AI section examines context engineering, human oversight, agent debugging, security, evaluation, and enterprise applications. Additional links span Spring, Quarkus, MongoDB, Kubernetes, Zig, and developer tooling, alongside an international calendar of Java conferences and community discussions on engineering culture, AI-assisted coding, Project Loom, Grails, Spring Security, and WSL support in JetBrains IDEs.
Oct 05, 2026
1,150 words in the original blog post.
RustRover and VS Code with rust-analyzer both support everyday Rust development but differ in architecture and workflow: VS Code is a lightweight, general-purpose editor whose Rust capabilities come from rust-analyzer and optional extensions, while RustRover is a dedicated JetBrains IDE built around its own Rust analysis engine and an integrated Cargo project model. Both offer code completion, diagnostics, navigation, refactoring, testing, and AI-assisted development, but VS Code emphasizes modular customization through extensions, tasks, terminal commands, and launch configurations, whereas RustRover combines Cargo tools, debugging, test results, inspections, project-wide refactoring, version control, and configurable run configurations in one environment. Rust debugging in VS Code generally requires an extension such as CodeLLDB, while RustRover includes it by default, and its indexed project model provides additional project and dependency views. The comparison suggests VS Code may suit developers who prefer a lighter, multi-language, highly configurable setup, while RustRover may appeal to those seeking tightly integrated project tooling for larger or longer-term Rust work, although it can feel heavier. Users switching to RustRover can import VS Code settings, use a familiar Cargo project, verify the Rust toolchain, wait for indexing, and test a known run or debug workflow; the article notes that its developer feedback came from a JetBrains study and is not necessarily representative of all Rust developers.
Oct 01, 2026
2,324 words in the original blog post.
JetBrains has opened an Early Access Program for JetBrains Air, an agentic-development interface available as a Marketplace plugin or in JetBrains IDE 2026.3 EAP builds. Air is not an AI provider or preinstalled agent suite; instead, it connects developers with existing agents and subscriptions, including Codex, Gemini, GitHub Copilot, Claude Agent, Junie, Cursor, and other ACP-compatible tools. Designed for coordinating multiple agents in parallel, it organizes work into sessions that show activity, costs, changed files, and commits across projects, while offering IDE-based code review, debugging, profiling, database exploration, and semantic search capabilities to improve agent workflows. Developers can launch sessions with attached IDE context, use editor tabs, terminal interfaces, or graphical chat, run work locally or in the cloud, and isolate parallel changes through temporary worktrees. JetBrains says data is not sent anywhere unless an agent or AI plan is used, third-party provider data goes directly to that provider, and Air can be disabled as a plugin. AI Assistant remains supported separately, while JetBrains expects Air to become its primary agentic workflow experience and is offering limited free Junie Lite runs for everyday coding tasks.
Oct 01, 2026
1,438 words in the original blog post.