- ChatGPT Canvas is a new way to interact with the AI chatbot. It offers a separate window to make targeted edits and collaborate with the tool further.
- It works with the GPT-4o model and you can refine text and code outputs seamlessly.
- ChataGPT Canvas also offers some shortcuts to quickly make adjustments to the generated output.
OpenAI has added a new feature called ‘Canvas’ to ChatGPT. It works with the GPT-4o model and aims to close the gap between AI and human interaction. ChatGPT Canvas offers a new interface where users can refine text or code output in a separate window, besides the chat interface.
ChatGPT Canvas allows you to highlight a certain part of the output and make changes then and there. Basically, you can select a portion of the output and give inline feedback, and Canvas makes targeted edits keeping the whole project in mind. Earlier, in a criticism of the chatbot interface, I had written that interaction with an AI system should not die in the four corners of a text chatbox.
With ChatGPT Canvas, OpenAI is addressing the limitation of the chatbot interface. So far, we have seen that chatbots are largely the de-facto way of interacting with AI systems. Anthropic recently introduced Claude Artifacts which offers a new way to interact with and visualize AI-generated output.
Similar to that, ChatGPT Canvas lets you polish your work without back-and-forth interaction with the chatbot. You can make direct edits in Canvas, like a copy editor or code reviewer. OpenAI has included some pre-made writing shortcuts for Canvas such as inline editing suggestions, document length adjustment, reading level customization based on audience expertise, grammar checking, and emoji addition.
As for code, Canvas shortcuts include reviewing code, adding logs and comments, detecting and fixing bugs, and translating the code to a different language (JavaScript, TypeScript, Python, Java, C++, or PHP).
ChatGPT Canvas is currently in beta and already available to ChatGPT Plus and Team users. OpenAI says ChatGPT free users will get Canvas access when it’s out of beta. Here is how to use the ChatGPT Canvas feature right now.
How to Use ChatGPT Canvas
- Open chatgpt.com (visit) and choose ‘ChatGPT 4o with canvas’ in the drop-down menu. Make sure you are subscribed to ChatGPT Plus.
- Next, enter the prompt on what you want to create. A separate window will open up with the generated output.
- Now, in the Canvas window, you can select a portion of the text and enter your suggestion by clicking on ‘Ask ChatGPT’. ChatGPT will make changes then and there. You can also choose to format the text.
- In the bottom-left corner, you will find Canvas shortcuts to refine your text further. For example, I changed the reading level to high school level. You can polish the text further and add emojis as well.
- While working with code, you can make similar inline editing in the code itself, just like Cursor.
- Further, you can access Canvas shortcuts and add comments and logs, fix bugs, and port the code to a different language.
- By the way, in the standard GPT-4o interface, you can write “use canvas” in your prompt to trigger the Canvas window without changing the model from the drop-down menu.
So this is how you can use ChatGPT Canvas to reiterate your work and collaborate to make the final output better and more readable. If you have any questions, let us know in the comments below.
