How we label AI content

For photos, we distinguish between standard editing and a visualisation that changes what is actually in the property. Only the second kind gets a label.

• Light, color, sharpness, straightened walls, cropping and zoom.
• Tidying up: clutter, personal items, rubbish, cars parked in front of the house, a washed facade, a cleaned pool, a mowed lawn.
• Small styling touches: table settings, cushions, towels, small decorations.
• Atmosphere: golden hour, dusk, sky, lights turned on.
• Blurring faces and car license plates.

• Adding furniture to a room.
• Removing furniture and clearing a room.
• Redoing the interior.
• Renovation visualisation: new kitchen, bathroom, floors, facade, garden.
• Inserting an object into the photo.

Detail shots, the drone shot, and the 3D floor plan are not cut out of your photo. Our system draws them from scratch. The subject is real, but the shot itself was not created that way in a camera. These shots do not carry a visible label, because they do not change what is actually in the property. The information that a computer created them is written directly into the file.

A label is not an admission of a defect. It tells the buyer that what they see in the photo is a proposal, not the property's current state. Photos that only change the quality of the shot and not the property itself do not need a label.

We also write an invisible piece of information into every photo, noting that our system worked on it. It is not visible in the photo, it does not change its quality, and it lets social networks and search engines recognize its origin too.