Brand, visualisation & web
Launching a boat before it touches the water
Conde Racing Lab had a fixed date to reveal Project33, a new foiling catamaran, before the boat could be photographed. We created its identity and launch website, and built a parametric 3D model of the boat with AI-assisted scripting, so every render and the launch film came from one consistent, believable geometry.
The challenge
A boat launch usually starts with photographs, and there was no finished boat to photograph. The identity, imagery and website all had to come together in the right order, against the clock. And the visuals had to convince people who know foiling boats inside out: the geometry had to be believable, not just impressive.
What we did
- Created the Project33 visual identity and a practical design system for the boat, sails, digital channels and launch material.
- Built a parametric 3D model of the whole boat with AI-assisted scripting, starting from the design team’s hull, beam and foil geometry.
- Produced photorealistic images and film from that one model, and a two-stage website: a pre-launch page, then the full reveal site.
Approach
1. Give the project its identity
We designed the Project33 mark and visual language around the character of the boat: technical, fast and distinctive. The identity was prepared in formats for everything from the hull and sails to the website and small digital applications, with a guide to keep future use consistent.

2. Build the boat in code
The design team supplied CAD geometry for the hulls, beams and foils. The rig, sails, rudders and deck were still to be modelled, so we built them as a parametric model, with AI assistance in writing the scripts. Every dimension and design reference stays explicit and editable in the code. It is a visualisation model, built for the camera, not manufacturing CAD.

Behind the model: AI-assisted parametric developmentThe code that builds the mainsail, beside the complete boat
Each part of the boat is a function of a few named dimensions. The mainsail, for example, starts from the boom and mast-head positions, the foot length and a target area: the luff follows the mast, and the script solves for the roach, the outward curve of the leech, that gives the required area. Change a dimension and the sail is rebuilt to match.
def mainsail():
tack_z = BOOM_Z + BOOM_D / 2
height = MAIN_HEAD_Z - tack_z
foot = BOOM_LEN - 100
def z(h):
return tack_z + h * height
def luff(h): # up the mast's trailing edge
return V(mast_te_x(z(h)), 0, z(h))
def build(roach):
# straight foot-to-head, plus roach
def leech(h):
w = (foot * (1 - h) + MAIN_HEAD_W * h
+ roach * math.sin(math.pi * h))
return V(mast_te_x(z(h)) + w, 0,
z(h) - BOOM_DROP * (1 - h))
return sail(luff, leech, MAIN_CAMBER, n=14)
# solve for the roach that gives the area
return fit(build, 0.0, 4000.0, MAIN_AREA)
3D model3. Turn one model into every image, and the site
We staged the model in realistic sailing positions and built water, sky, spray and lighting around it. The same geometry fed the stills and the launch film, so the boat stayed consistent from one camera to the next. Visualisations were presented as renders, never as photographs, and images with visible artefacts were left out. The website went live in two stages: a pre-launch page first, then the full site after the reveal.
3D model
Render
Outcome
Conde Racing Lab had a complete launch package, a recognisable identity, a finished website and a set of visuals, ready before the boat could be photographed. Because every image came from one parametric model, design updates could be carried through the renders without rebuilding each view. On 27 October 2026, Project33 was revealed at Costa Nova and the real boat met the water.


Facing a similar challenge?
A first conversation takes about 30 minutes and costs nothing.