Analyze ANY Website Visually - Without Exporting 10 Screenshots

The feature I think Poppy should have is…

A simple way to add a website to the board as an element that Poppy’s AI can visually see, not just read as extracted text.

By “see,” I mean understanding the website as it actually appears:

• The layout

• The design

• The visual hierarchy

• The images

• The sections

• The positioning of different elements

• The relationship between the text and visuals

The AI chat connected to that element should be able to access this visual version of the website, whether the element is by itself or inside a group.

Ideally, I could paste a link or import the website once, connect it to a chat, and ask questions like:

• “Analyze the structure of this landing page.”

• “Why does this design feel trustworthy?”

• “Describe the visual hierarchy.”

• “Create a page with a similar structure.”

• “Compare the layouts of these three websites.”

• “Which sections should I recreate for my own offer?”

And Poppy would be able to analyze what is actually on the screen.

The current problem

As far as I understand, the only reliable way to make the AI visually see a website is to:

1. Export the website as an image.

2. Split it into multiple images if it is taller than 8,000 pixels.

3. Upload those images directly inside the chat.

This works, but it creates a lot of unnecessary manual work.

Adding the same image to the board does not appear to work in the same way. Instead, the image is automatically converted into a written description, and the AI chat can only access that processed description, not the actual image.

A PDF does not solve the problem either.

I can export an entire website as a searchable, one-page PDF that looks exactly like the original page. But when I add that PDF to the board, the AI can read its text without visually seeing the page.

I also cannot upload the PDF directly inside the chat because that file format is not currently accepted there.

So the workflow becomes:

Export → split → upload multiple images into the chat → repeat for every website.

This would help me because…

Websites are not just collections of text.

Their layout, visual hierarchy, spacing, imagery, and structure are often the most important parts—especially when researching landing pages, marketing, design, branding, or content presentation.

I want to be able to add a website to the board with one simple action and let Poppy analyze both:

1. What the website says

2. What the website visually looks like

The ideal version would allow Poppy to visually render and analyze a website directly from its link.

But even allowing the AI to visually access an imported full-page PDF or board image would make this workflow much faster.

The main goal is simple:

Let me add a website once and allow any connected AI chat to actually see and understand it—without manually exporting, splitting, and uploading multiple screenshots.