A library you can actually search

Asset Tagging

Tags generated from what the model looks like, from a controlled vocabulary — so “vehicle” always means vehicle and a filter finds every one of them.

What it does

Looks at the model, from four angles

Views are captured from the viewer and read together, because what is hidden from the front is often obvious from behind.

A controlled vocabulary

Category, style, theme, materials and colours come from fixed lists, which is what makes tags searchable rather than a pile of synonyms.

You approve them

Suggestions arrive selected; tap to drop any that do not fit, then add the rest. Nothing is written until you say so.

And add your own

A free-text field for the things a vocabulary cannot express — a client name, a project, an internal label.

How it works

  1. Open a modelTagging lives with the tags, in the Inspect panel.
  2. Generate tagsFour views are captured and read. The polygon count is supplied too, because whether a mesh is low-poly is not visible in a render.
  3. ReviewDrop what does not fit. Tags already on the model are not offered again.
  4. Add and remove freelyType your own, or remove a saved tag with its ✕.

What people use it for

  • Making a growing library findable by something other than an id
  • Consistent metadata across assets made by different people
  • Filtering a library by material, style or theme
  • Labelling uploaded models that arrived with machine names

What it cannot do

  • It reads pictures, so it describes appearance rather than intent
  • The vocabulary is fixed — your own words go in the free-text field
  • Uses AI tokens, and the cost is reported after the run rather than promised before it

The details

Input
Four viewer frames, plus the polygon count
Facets
Category, style, theme, materials, colours, subjects
Review
Required — nothing is saved unapproved
Meter
AI tokens