OpenAI Launches Canvas, a New Coding Interface in ChatGPT

OpenAI has launched Canvas, a new interface for writing and coding that expands ChatGPT’s functionality beyond simple conversation. The interface allows users to work on writing and coding projects side by side with ChatGPT, offering real-time edits, feedback, and suggestions.

Available globally to ChatGPT Plus and Team users, Canvas will also roll out to Enterprise and Edu users next week, with plans for Free users once the beta phase concludes. Canvas is integrated with GPT-4o and can be manually selected in the model picker. 

With the introduction of Canvas, ChatGPT is now able to better understand the context of user projects. Users can highlight specific sections, allowing ChatGPT to focus on targeted areas and provide inline feedback and suggestions, similar to the role of a copy editor or code reviewer.

Users maintain full control of their projects in Canvas, with the ability to directly edit text or code. A menu of shortcuts enables quick actions such as adjusting writing length, debugging code, and performing other tasks. Users can also restore previous versions of their work through a back button feature in the Canvas interface.

The coding shortcuts available in Canvas include several helpful features for developers. ChatGPT can review code and provide inline suggestions to improve it, making the coding process more efficient. It can also insert print statements, or logs, to help with debugging and understanding the code. 

Additionally, ChatGPT can add comments to make the code easier to follow and detect and rewrite problematic sections to fix bugs. Another key feature is the ability to port code to different languages, including JavaScript, TypeScript, Python, Java, C++, or PHP.

“Did OpenAI just kill Cursor (and all the other VSCode forks?),” posted a user on X. 

“OpenAI could compete with Canvas CursorAI and other AI editors,” posted another user. 

The system automatically opens Canvas when it detects complex scenarios and can be triggered manually by including “use canvas” in a prompt. 

OpenAI said that Canvas improves decision-making accuracy for when to trigger writing and coding tasks, reaching 83% and 94%, respectively, compared to baseline models.

OpenAI has trained GPT-4o to perform targeted edits based on user input, improving efficiency in writing and coding projects. The company intends to continue refining Canvas based on user feedback during the beta phase.