ChatGPT, the AI-powered chatbot platform developed by OpenAI, has now gained the ability to directly edit code, but this feature is currently exclusive to macOS users.
The latest update to the ChatGPT app for macOS enables it to edit code in several supported developer tools, including Xcode, VS Code, and JetBrains. Additionally, users have the option to enable an “auto-apply” mode, which allows ChatGPT to make edits without requiring further user interaction.
As of Thursday, subscribers to ChatGPT Plus, Pro, and Team can utilize the code editing feature by updating their macOS app. OpenAI has announced that this feature will be rolled out to Enterprise, Edu, and free users in the coming week.
According to a post on X by Alexander Embiricos, a product staff member at OpenAI working on desktop software, the ChatGPT app for Windows will also receive the direct code editing feature “soon.” Embiricos’ post provides more information about the update. Furthermore, the ChatGPT app for Windows is expected to get this feature in the near future.
ChatGPT for macOS now has the capability to edit code directly in IDEs, available to Plus, Pro, and Team users. pic.twitter.com/WPB2RMP0tj
— OpenAI Developers (@OpenAIDevs) March 6, 2025
The direct code editing feature builds upon OpenAI’s “work with apps” capability, which was launched in beta in November 2024. This feature allows the ChatGPT app for macOS to read code in several developer-focused coding environments, reducing the need for users to copy and paste code into ChatGPT. For more information, see the original announcement.
With the introduction of direct code editing, ChatGPT now competes more directly with popular AI coding tools like Cursor and GitHub Copilot. OpenAI is reportedly planning to launch a dedicated product to support software engineering in the coming months.
The adoption of AI coding assistants is on the rise, with the majority of respondents in GitHub’s latest poll indicating that they have adopted AI tools in some form. Jared Friedman, a partner at Y Combinator, recently stated that a quarter of YC’s W25 startup batch have 95% of their codebases generated by AI.
However, there are several security, copyright, and reliability risks associated with AI-powered assistive coding tools. A survey by Harness found that the majority of developers spend more time debugging AI-generated code and addressing security vulnerabilities compared to human-written code. Additionally, a report by Google discovered that while AI can accelerate code reviews and improve documentation, it may compromise delivery stability.
