Fresh off the news of its record-breaking $6.6 billion funding round, OpenAI is updating its signature AI app ChatGPT with a new feature – Canvas – that allows users to see only select portions of chatbot output in one , directly edit and easy to change panel view side by side.

The feature, based on OpenAI’s GPT-4o model, can also make suggestions and implement changes to the answers in the right pane without having to issue an entirely new answer. It appears to be a direct challenge to AI startup Anthropic’s Artifacts feature for its chatbot Claude, which launched in June 2024. It also provides a side panel view for viewing and running simple Python programs that update based on the user’s prompts.

Canvas is rolling out to ChatGPT Plus and Teams users, with Enterprise and Edu tier subscribers following in a few days. After the beta, OpenAI plans to make the feature available to all ChatGPT users.

Social network users

What ChatGPT Canvas is good for

Daniel Levine, product manager for Canvas, told VentureBeat in an interview that the vertical, top-down chat window is sometimes too limited for some of the most common use cases.

“We know that many people use ChatGPT for writing and coding. Those are two of the most common use cases we see,” Levine said. “But the chat interface is a bit limiting, especially for projects that require revisions or edits. There’s a lot of back and forth and comparing changes is difficult, so that’s where Canvas comes in.”

Levine talks about asking ChatGPT to edit its responses. If users ask ChatGPT without Canvas to create a draft email and feel the response needs to be longer, funnier, or friendlier, they will have to ask ChatGPT again and rewords will appear in the same conversation.

Sometimes the design changes drastically; Sometimes the change is subtle, but often you have to scroll back and forth to double-check what has changed.

OpenAI hopes Canvas will make the process easier. Users can of course revisit ChatGPT, but if they just want to change a few words, they don’t have to copy the text, which is still a first draft, into another document and edit it themselves.

How to use Canvas

Canvas opens after users toggle the model selection, where they typically select which version of OpenAI models to use. ChatGPT can also detect when Canvas should be opened or when the user sends a “Use Canvas” prompt. They can then ask ChatGPT to either write something or generate code. For example, if a user wants to write an email to a colleague, they send the corresponding request to ChatGPT and open the Canvas window with the just generated text.

Users can still ask ChatGPT to refine the text that appears on the Canvas screen. People can also make changes or give instructions to the chatbot directly in Canvas by highlighting text. In each Canvas window, there are a number of shortcuts that users can click that allow them to adjust text length, reading level, add emojis, or even add finishing touches.

ChatGPT can also provide suggestions for the text that appears in text boxes on Canvas, similar to comments in Google Docs. It can also be translated into supported languages.

The canvas looks different depending on the task. The writing canvas looks like a Word document, while the coding canvas contains line numbers to make code editing easier. The key combinations for coding will also be different. Users can review code, fix errors, add logs and comments, and port it to another programming language through Canvas.

Directly competes with Anthropic Claude’s artifact feature

Canvas’s separate window and model collaboration is reminiscent of another chatbot with a window where users can clearly see any changes made through new prompts: Anthropic’s Claude Artifacts.

VentureBeat’s Michael Nunez reported that Claude Artifacts makes accessible and easy-to-understand interfaces a key feature of chatbots, calling it “the most important AI feature of this year.” Unlike Canvas, Artifacts is already generally available to all Claude Chat users.

Artifacts also allow users to see what their generated code looks like, as users can ask Claude to write code, edit it, and then see the fruits of the labor, such as a prototype of a website or game. Canvas simply shows users the generated code and the changes made to it.

The battle over the new interface, led by OpenAI’s Canvas and Anthropic’s Artifacts, points to a problem that smaller, third-party AI applications have been trying to solve: how to make chatbots easier to read and use.

Apps like Hyperwrite, Jasper, and JotBot all generate and offer text editing. Many other software programs offer similar code and text editing capabilities with the idea that users don’t have to leave the chat window to make changes to their work.

Over time, features like Canvas and even artifacts could become commonplace as more people want to work more efficiently and collaboratively with chatbots.

“We believe that collaborative work is an important part of the workplace,” Levine said. “So we’re taking a first step here.”