Anvekra.
All workCase study · 3D · Configurator · WebGL

Gothic Victorian Carriage

A 100% code-generated 3D carriage with a bespoke configurator — finishes, drive mode and an exploded view.

Discipline
Frontend & Motion
Year
2026
Context
Design showcase
Live site
carriage.anvekra.com

(01)Overview

A gothic Victorian carriage built entirely from code — no hand modelling — and presented as an interactive product configurator. A Python pipeline generates the geometry and PBR textures in Blender; the browser streams the result into a Three.js scene where visitors restyle the coach, open its doors, drive it down a cobbled road or explode it into parts.

(02)What we built

  1. 1. Every part — barrel-vault roof, gothic arches, 12-spoke wheels, leaf springs — generated procedurally in Python.
  2. 2. Live finishes for wood, wheels, metalwork, glass, seating and emblems, plus four curated themes.
  3. 3. Drive mode on a procedural cobblestone road, with real-time suspension that reacts to the stones.
  4. 4. Open doors and windows, toggle lanterns, wireframe and an exploded view of the whole build.
  5. 5. Annotated callouts that track the model, and technical elevation drawings for side, front and rear.
  6. 6. Ships as a static site: plain ES modules and three.js from a CDN, no bundler.

(04)Engineering highlights

01

Model as code

The carriage, its materials and its textures are produced by a headless Blender script, so any proportion or detail can be regenerated rather than re-modelled.

02

Suspension kinematics

A multi-body solver moves wheels, springs and body independently so the coach rides over the cobbles believably.

03

Zero-allocation callouts

The SVG leader lines are updated every frame without creating garbage, keeping the render loop smooth.

04

Performance monitor

A built-in frame-time monitor keeps quality in check on lower-powered devices.

(05)Inside the build

Screenshots from the running system — select any image to enlarge.

Next case studyFutureRideA three-vehicle 3D showcase where the model really sits in front of the type — 83 MB of assets shipped in 6.9 MB.Read case study
futureride.anvekra.com
FutureRide preview
Keep exploring

More work

View all work

Want something like this?

Tell us about your product or platform — we'll get back to you within one business day.

Get in touch