I’d love to see a lightweight shared visual workspace or whiteboard built directly into ChatGPT conversations.
The idea isn’t simply to add a drawing app. What I’d really like is a shared visual communication space between the user and ChatGPT.
Imagine something similar to the side panel that opens when viewing generated images. While continuing the same conversation, the user could open a blank visual workspace and quickly sketch with a mouse, stylus, or touchscreen.
ChatGPT could see the current state of that workspace as part of the conversation context.
This would be especially useful for things that are awkward or inefficient to describe purely with text, such as:
- sketching a rough UI/UX wireframe and discussing the layout together
- arranging a character reference sheet
- indicating image composition or camera direction
- roughly sketching poses
- drawing arrows for movement or spatial relationships
- explaining 3D structures or clothing layers
- diagrams, game mechanics, level layouts, and other spatial ideas
- annotating areas that should be changed or refined
The drawings wouldn’t need to be polished. Simple boxes, circles, arrows, stick figures, handwritten notes, and rough lines are often enough to communicate an idea much faster than several paragraphs of text.
Ideally, with the user’s permission, ChatGPT could also add simple annotations, shapes, arrows, or text to the same workspace. The user could then draw over or modify those suggestions, allowing both sides to iteratively refine the idea.
For example, when designing a UI, I could quickly sketch where I imagine the character, inventory, buttons, and information panels should go. ChatGPT could immediately understand the spatial relationship and discuss usability, hierarchy, or alternative layouts with me without requiring us to translate the entire visual idea into words first.
Why keep it inside ChatGPT?
I experimented with using an external visual collaboration tool such as Figma/FigJam for this purpose.
It is powerful, but moving between ChatGPT and a separate workspace adds friction. For a quick sketch, it can actually be faster to draw something in a basic drawing program, take a screenshot, and paste it into the conversation.
The value of this feature would therefore come from keeping the visual workspace inside the current conversation and its context.
The ideal interaction would be almost frictionless:
Conversation → quick sketch → ChatGPT sees it → discussion → refinement → creation
This could also be particularly useful for image generation. Instead of trying to describe every spatial detail through increasingly complicated prompts, the user and ChatGPT could first establish the composition together visually, then generate the image once they agree on the idea.
More broadly, I don’t really think of this as a “whiteboard feature.”
I think of it as adding a visual language to the conversation.
Text is excellent for communicating concepts, but sometimes a five-second sketch communicates spatial intent better than five paragraphs.
A shared visual context could make collaboration with ChatGPT significantly faster, clearer, and more natural.
Would anyone else find something like this useful? I’d especially be interested in how people might use it for UI/UX design, image creation, diagrams, spatial reasoning, education, or other collaborative work with ChatGPT.