SwannBuddy 4K AI doorbell/surveillance camera in testing – New dimensions with artificial intelligence

The SwannBuddy 4K is a combination of a doorbell and a video camera and promises to bring artificial intelligence into the doorbell and even act as your own lobby attendant. But how far does artificial intelligence go and how does automation actually work?

The SwannBuddy 4K is a combination of a doorbell and a video camera. So it's not really anything new under the sun: there are plenty of these on the market.

In this case, however, the difference is sought primarily through technical solutions and the promise of extensive AI functionality. 4K resolution, vertical viewing angle and local storage without mandatory cloud payment are the manufacturer's key arguments. There are various automation and AI functions included, which are intended to bring more intelligence to the device compared to a traditional video doorbell.

Actually, it's more a combination of a doorbell, surveillance camera, and artificial intelligence than a traditional video doorbell. This also explains the rather high price of 169 euros. In this case, however, without the internal display screen, which is replaced by a smartphone. Which of course also offers functionality via the network anywhere.

The camera has a resolution of 4K (3840 × 2160), but in practice the more important solution is the aspect ratio and angle of view. Swann uses a 1:1 image and a vertical view, which covers the door area from top to bottom. This allows both the visitor and, for example, packages left on the floor to be seen in the same image.

The image quality is detailed, and in good lighting the resolution is clearly enough to recognize faces and smaller details. However, the wide angle of view brings with it a slight distortion, especially at close range, which is typical for this type of solution.

In dim and dark conditions, the image is based on infrared illumination. It is sufficient for basic use, but the image resolution and contrast are expectedly reduced compared to daylight. Overall, the focus is more on practical coverage than on resolution alone.

Image storage is implemented locally on a microSD card, and continuous use does not require a cloud service. This is a structural difference from many competitors, where storage is tied to a monthly cloud fee. Cloud storage is of course available as an additional option.

Motion detection is implemented using a PIR sensor, which reacts to heat and movement. The goal is to limit false alarms and focus recording on real events. The detection works in conjunction with software-based analysis, which aims to distinguish, for example, human movements from other movements.

The voice connection is said to be two-way, but the implementation is not so-called full-duplex. In practice, the connection is half-duplex, meaning only one party can speak at a time. Speaking also requires pressing a separate button in the application, which makes the conversation more intermittent than in a continuous voice connection. The user can therefore speak to the visitor from the application and alternatively use ready-made voice messages or automatic responses. These are controlled by the SwannShield function, which is an additional AI-based feature from the manufacturer.

In terms of power supply, the device operates either on battery or connected to a fixed power supply (requires a separate transformer). The battery life is promised to be up to several months, but in practice the duration clearly depends on use: the more motion detections, recordings and live viewing, the faster the battery will drain. Cold conditions shorten the operating time even further. Otherwise, the device is promised a minimum operating temperature of “only” -10 degrees, which is a clear limitation in at least part of Finland, even though winters are warming up.