Jesse Moffett.
01 / Self-directed concept

KIN 01

A concept product site for a fictional household robot, told as a scroll-driven film of one ordinary family’s evening, night and morning.

Art DirectionWeb DesignMotionAI Film
Launch the live site

About the project

Problem

Make a home robot feel like an ordinary, trustworthy purchase, not science fiction

Humanoid home robots are becoming real consumer products, so KIN is pitched as something a family finances, the way it would a car.

Solution

One feeling: you woke up and the house was already fine

Instead of a feature list, the page shows a family getting their time back, and lets the robot’s skills and safety follow as proof. The constraint I set: the world moves and the type stays still. Calm editorial type and glass panels sit on top while films play behind them.

Type
Self-initiated concept
No client
Role
Concept, art direction
UI and motion design
AI film direction
Built with
Claude Code
Three.js
Seedance, Omni
Timeline
Aug – Sep 2026

Note  KIN 01 is a fictional product. The robot, its price and its launch date are invented, and the reservation form stores nothing.

Live site  kin.jessemoffett.com

The design

One Evening to the Next Morning

The house, handled.

Home hero — the head follows your cursor
The KIN home page hero: the cream-and-copper robot on a warm dark ground beside the headline The house, handled., with the Reserve button and pricing below

Life with KIN

Film chapter with glass card
KIN drying a plate at a white kitchen counter, a glass card reading KIN handles the rest on the right and a portrait inset of the robot on the left

A little more time together

Morning — the card opens with the oven
Seen from inside the oven, KIN slides a dish of lasagna onto the rack, while a glass card and a family photo inset sit over the frame

Inside KIN

Abilities page — interactive 3D exhibit
A dark hairline 3D model of KIN's hand on a plinth, closed in a grip with its working joints lit amber, beside buttons for open, pinch, point and grip

How it was made

Process

Every film is AI-generated from approved reference images, then built into a hand-coded scroll experience

Lock the character first. I approved a small set of KIN reference images (full body, close-up, side views) and used them as the look for every shot, so the robot stays the same robot.

Direct each film in one sentence. Each scene is image-to-video: the reference frame carries the look, and the prompt only says what happens.

Guide the motion when words can’t. For precise moves, like fingers turning a ring, I built a grey Blender motion guide and fed it to the model as the motion reference.

Build the experience by hand. A custom scroll engine pins each scene and reveals the next with a curved wipe, one gesture per chapter; films play once and hold. The 3D exhibits on the Abilities page are built in Three.js.