Device frame or plain screenshot: decide per placement

There is no universal answer, and the people who argue for one are usually generalising from a single placement. A frame adds context and costs screen area. Which trade-off wins depends entirely on where the image will be seen and how big it will be.

Frames win where context is missing

In a store listing, on a landing page hero, in a social post — anywhere the viewer has not yet worked out what kind of product this is — a device frame answers "what am I looking at" instantly. It also implies platform: a phone body says mobile app, a laptop says web. That is a lot of communication for the cost of some margin, and it is why store listings almost universally use frames.

Plain wins where space is scarce or context is known

In documentation, in a feature section halfway down a page, in a changelog, in an email — the reader already knows it is your app. A frame there just shrinks the interface by 30 percent for no information gain. The same applies to any image displayed small: below roughly 400 px wide, the frame eats the content. Crop to the relevant part of the interface instead, and let the surrounding page supply the context.

A middle option: partial frames and crops

You do not have to choose the whole device. A phone cropped at the bottom edge of the canvas keeps the recognisable top curve and status area while giving the interface more room. A screen shown at an angle with the body bleeding off the frame does the same. These half-measures are common in good listings precisely because they capture most of the context benefit at a fraction of the space cost.