Digital lighting showroom
A 74,722‑item catalog, search for a fixture by photo, virtual try‑on in the buyer’s interior and specs for designers.
Task
A lighting store needed a catalog where buyers quickly find a model among tens of thousands of products and interior designers get project data straight away.
Searching by name is awkward with a range this large: buyers often do not know what the fixture they liked is called, and scrolling through tens of thousands of product cards takes forever.
Solution
A 74,722‑item catalog
The store’s entire range in one catalog, with photos and specifications for every fixture.
Search by photo
The buyer uploads a photo of a fixture and the CLIP neural network finds similar models in the catalog. No name or SKU needed.
Virtual try‑on
The buyer uploads a photo of their room and sees how the chosen fixture will look in it before buying.
Specifications for designers
The site builds a specification from the selected fixtures, so the interior designer gets ready‑to‑use project data.
Own infrastructure
The site runs on its own Ubuntu server with LiteSpeed, services are packaged in Docker and media is processed with FFmpeg.
Result
74,722
items in the catalog
By photo
search for similar fixtures
In your room
try‑on on the buyer’s photo
Buyers find a fixture by photo and see it in their own room right away, while designers get a specification without manual work.
Palette
Click a colour to copy its HEX
Technologies
Django
Python
PHP
CLIP
Docker
SQLite
Ubuntu
LiteSpeed
FFmpeg21 services in one window: competitor analysis, a prioritised work list, AI content production, plans and client reports. Calls are transcribed locally, with no cloud.
View case
