Official Kotlin Support for VS Code Now in Alpha — Powered by IntelliJ's Language Server
JetBrains has released the Alpha version of the official Kotlin extension for Visual Studio Code, announced at KotlinConf 2026. The extension, called Kotlin by JetBrains, connects VS Code to the Kotlin Language Server — which is built on IntelliJ IDEA's code-insight infrastructure and the Kotlin plugin implementation. This means the same analysis engine used in IntelliJ IDEA now powers Kotlin features in VS Code.
What's in the Alpha
- Code completion – context-aware suggestions based on Kotlin type system
- Diagnostics – real-time error and warning highlighting
- Navigation – go to definition, find references, symbol search
- Quick-fixes – actionable suggestions for common issues
- Formatting – code style consistent with the Kotlin coding conventions
- Project import – automatically recognizes Gradle, Maven, and other Kotlin project structures
To get started, install the Kotlin by JetBrains extension from the VS Marketplace, open a Kotlin project, and start coding. No additional configuration required — the language server starts automatically.
Context: Part of a Broader Kotlin Tooling Push
This release follows the Java to Kotlin Converter extension for VS Code released earlier this year. Developers using editors that support the Language Server Protocol can also try the standalone Kotlin Language Server, though setup requires manual configuration and an editor with pull-based diagnostics support.
Who It's For
Kotlin developers who prefer VS Code over IntelliJ IDEA or Android Studio. IntelliJ IDEA remains the most complete environment, but this extension closes the gap for VS Code users.
Give Feedback
Since this is an Alpha release, JetBrains is actively seeking feedback. Report issues or share suggestions on the GitHub repository.
📖 Read the full source: HN AI Agents
👀 See Also
Claude Command Center: Open-Source Dashboard for Claude Code Analytics
Claude Command Center is a local dashboard that reads your ~/.claude/ directory to display Claude Code session data, costs, and MCP server configurations. Built entirely using Claude Code with an Express backend and React frontend, it requires zero configuration and runs locally with no cloud or telemetry.
LM Studio parser bugs break Qwen3.5 tool calling and reasoning
LM Studio's server parser has three interacting bugs that silently break tool calling, corrupt reasoning output, and make models appear worse than they are. The issues affect reasoning models like Qwen3.5 and DeepSeek-R1, with one bug reported over a year ago still unresolved.
Why Deterministic Workflows Outperform AI-Driven Orchestration for Agent Systems
A developer with a year of experience building agent systems shares that AI-driven orchestration failed reliably due to non-deterministic routing, compounding errors, cost explosion, and impossible debugging. Switching to deterministic workflows with code-based orchestration eliminated orchestration failures.
Unlocking Proactivity: A Deep Dive into Clawbot Innovations from the Community
Discover how enthusiasts are enhancing their Clawbot's proactivity through inventive strategies and community-driven insights. A look at discussions and revelations from r/openclaw.