On Tuesday, OpenAI officially launched its Canvas tool, marking the fourth day of a planned 12-day release schedule. Initially unveiled in October, Canvas functions as a sandbox-style interface that enables users to make inline modifications, formatting, and adjustments to both text and code. Users have the flexibility to edit content directly or highlight specific sections to request changes from the AI. OpenAI has noted that Canvas is particularly effective for creative writing and coding tasks. This tool is now accessible to all ChatGPT users around the globe.
Following the rollout of the complete o1 AI model and the Sora video generation model, OpenAI has expanded the availability of the Canvas tool beyond its preview phase. This feature is built upon the advanced GPT-4o model, which is required for access. It’s important for users to note that those using the free tier of ChatGPT have limited access to this AI model.
The Canvas tool in ChatGPT
Canvas serves as an interactive collaborative window, allowing both users and AI to engage simultaneously in various projects. Traditionally, ChatGPT responds to prompts, and if users find the output unsatisfactory, they can offer additional details or submit new prompts in a trial-and-error approach. However, this method can be time-consuming and may necessitate several attempts for results that meet their expectations. This can be particularly challenging when debugging code, where guiding the AI to identify and correct specific issues can prove complex.
With the launch of Canvas, ChatGPT users can now locate a new “View tools” icon at the bottom of the text field, situated between attachment and web search options. The “View tools” menu includes both the Picture mode, which utilizes DALL-E, and the Canvas mode. Upon selecting Canvas and submitting a prompt, the chatbot automatically opens a pop-up window to generate the response.
After the response is generated, users have the option to manually add formatting or make edits. They can also highlight specific words or sections of text and input prompts to refine only those areas. This detailed editing capability is designed to assist users in achieving precise outcomes and addressing minor issues without navigating away from the ChatGPT interface.
