Introducing the all-new interactive interface ChatGPT Canvas, the ultimate tool for AI programming and writing

OpenAI has just raised funds and is eager to prove itself. Early this morning, OpenAI announced the launch of a new interactive interface called Canvas, similar to Anthropic’s Artifacts, describing it as “a new way to use ChatGPT for writing and programming.” It is a major update to ChatGPT’s visual chatbot interface launched 2 years ago.

ChatGPT’s new interface mirrors Anthropic’s Artifacts, launched in June and the recently popular AI code editor Cursor.

Those who have tried Artifacts in Claude know that it can significantly enhance the output of large language models (LLMs), supporting the export of text files, code, web pages, SVGs, and more. The once-popular project “New Interpretations of Chinese” was based on Claude’s Artifacts feature. However, much to the disappointment of ChatGPT users, OpenAI had not followed suit for over three months since Artifacts’ release, leading some developers to create and release their own open-source versions.

Now, with OpenAI finally responding to user demands, people are enthusiastically showing their support. Some even joked that OpenAI, unable to beat Claude, decided to join the game instead.

What is Canvas

In the Canvas interface, you can collaborate with ChatGPT on writing and coding projects, moving beyond simple conversation. Canvas is a new interactive format and represents OpenAI’s first major visual interface update since ChatGPT’s release.

Canvas opens in a separate window, making it easier for users to collaborate with ChatGPT on projects. The Beta version of Canvas offers users a new way of working together: not only can you create through dialogue, but you also get to collaborate with ChatGPT as a partner to create and refine content.

Collaboration with ChatGPT

Chatting with ChatGPT has always been a convenient way to gather information, including for writing and coding. However, when you need to make immediate changes to the content, the dialogue format might feel limiting.

Canvas was created to solve this problem.

In this new interface, you can highlight specific parts to let ChatGPT know what to focus on, allowing it to understand your intentions more precisely. It’s like editing in context, where you can provide feedback and suggestions.

Moreover, you can easily make changes by directly editing the code or text. The project is under your control. Canvas also offers a quick-access menu that lets ChatGPT help you adjust text length, debug code, or perform other useful actions. If you want to go back to a previous version, one click is all it takes to restore it.

Canvas for Programming

Coding is an iterative process, but it can be hard to track changes through conversation alone. Canvas makes it easier to follow and understand the modifications ChatGPT makes. OpenAI has also promised to continue enhancing transparency in this editing process.

Programming Quick Actions Available in Canvas:

• Review Code: ChatGPT can provide inline suggestions for improving code.
• Add Logs: Insert print statements to help with debugging and understanding the code.
• Add Comments: Add comments to make the code easier to understand.
• Fix Bugs: Detect and rewrite problematic code to fix errors.
• Support for Multi-language Export: Convert code into JavaScript, TypeScript, Python, Java, C++, PHP, and other languages.

Training the Model as a Collaborative Partner

OpenAI’s research team has trained GPT-4o to work as a creative collaborator. It knows when to open Canvas, when to perform targeted edits, and when to rewrite entire sections. It also understands the broader context, allowing it to give precise feedback and suggestions and improving the model’s ability to generate high-quality comments. The result is a model that performs better in collaboration tasks like writing and coding.

Canvas Beta Version Access

Canvas is currently in Beta, available only for ChatGPT Plus and team users. Enterprise and educational users will get access next week, while free users will have to wait until Canvas is officially released.

For now, let’s stay tuned to see how Canvas measures up to Artifacts ans Cursors, and whether OpenAI can continue meeting user expectations after their recent funding.