Character Quality Tiers

Why characters cost more than objects

People, creatures, and pets are the hardest subject in 3D reconstruction: faces, fur, and organic shapes punish every shortcut, and the reconstruction engines that actually win on them cost more to run than the ones that handle a chair or a go-kart. So while objects start at $0.25, characters start at $0.75, and the price you pick buys a genuinely different engine, not the same one with a knob turned up.

The three tiers

• •Budget, $0.75: the cheapest character recipe. Fine for iteration, blockouts, and background NPCs, and a sensible first pass when you queue a lot of characters and plan to upgrade the keepers.
• •Standard, $1.00 (the default): what you get if you choose nothing. On our benchmark it held the subject's identity best across stylized characters, animals, and photoreal people, which is exactly the thing you care about when the model is supposed to be a specific someone. For most characters this is the tier to use.
• •Hero, $3.00: figurine-grade, the most detail we can produce. For the centrepiece model: the player character, the pet you are going to print, the one that gets the closeup. Not worth it for a crowd.

Each queued image shows its detected subject and price before you commit, and a character row gets a quality toggle right in the queue. See Wallet & Pricing for how billing works.

What counts as a character

Any living being photographed as itself: a person, a pet, any animal, a creature, a robot, or a cartoon or video-game character. Detection is automatic, and pets count as characters because they need the same engines a person does. The flip side saves you money: a photo of a manufactured figure, a plush toy, an action figure, a figurine, a statue, is an object, because what you photographed is the manufactured item, and it prices at object rates from $0.25. A real dog on the couch is a character; a stuffed toy dog on a shelf is an object. If the detected badge is wrong, click it and change it; the price follows the badge.

The pose in the photo is the pose in the model

Characters keep the exact pose from your photo: a dog curled on the couch comes out curled, a character mid-swing stays mid-swing. We treat this as identity, not styling, because a lying dog re-posed standing reads as a different dog. The trade-off is honest: a self-occluding pose hides some of the subject from the camera, so the hidden side is the model's best reconstruction rather than observed detail. A photo where the limbs are clear of the body gives the engines the most to work with.