Visual Snapshot - why we render listings instead of only reading DOM

Why not just read DOM?

A listing is more than text. Layout, image aspect, container heights, font sizes - all of that contributes to the customer's impression. If we only read the DOM, we optimize blind. After publishing the text suddenly overflows the card.

What we do

1. Playwright renders your listing in a headless Chrome.
2. We extract bounding boxes of all relevant elements (title

container, description block, images).

1. From that our optimization AI gets constraints:

- "Title may be max. 84 characters" - "First description paragraph has 280 px height - that's what the customer sees"

1. The AI only produces suggestions that fit those constraints.

Where this is visible

In the optimizer workflow you see:

• Left: your listing with snapshot
• Right: the AI proposal
• Both rendered in the same renderer - you see immediately if the

layout shifted