Codex in JetBrains IDEs

Quick Overview

JetBrains has integrated OpenAI's Codex AI into its IDEs, allowing developers to use AI assistance for coding tasks like debugging, feature implementation, and localizing applications, demonstrated by successfully adding Spanish localization to the multiplatform KotlinConf App.

Key Points: Codex is now integrated directly within JetBrains IDEs (IntelliJ IDEA, Rider, WebStorm, PyCharm) for AI-assisted coding, requiring a ChatGPT subscription or API key for activation. The demonstration uses the KotlinConf App, a Kotlin Multiplatform project, to showcase Codex's ability to assist with debugging and feature implementation across mobile, web, and desktop targets. Codex successfully analyzed a build failure error related to missing SVG implementation for iOS, investigated relevant files (like Svg.ios.kt and build.gradle.kts), and suggested/applied a fix by adding an iOS actual for SVG. The AI agent was prompted to add Spanish localization to the KotlinConf App, following a detailed set of requirements including platform-specific logic, resource setup, and adding a visible locale badge. The agent performed extensive research across multiple project files (e.g., Platform.kt, strings.xml, BaseScreens.kt) in parallel with the developer's work, demonstrating efficient context awareness. After applying changes, the agent successfully compiled the project for iOS Simulator (Arm64) and verified that the app launched with Spanish localization displayed on the home screen. The Agent mode within Codex allows the AI to edit and run commands directly in the IDE, requiring user approval for potentially sensitive actions like running shell commands.

Context: This video features two presenters from JetBrains demonstrating the integration of OpenAI's Codex AI within various JetBrains Integrated Development Environments (IDEs) such as IntelliJ IDEA, Rider, WebStorm, and PyCharm. The core concept being showcased is how developers can leverage this AI tool for complex, context-aware coding assistance, debugging build issues, and implementing new features directly within their development environment using either a subscription or API key.