Tagger — tag & tidy your asset library

Tagger turns a pile of mismatched asset packs into one clean, named, engine-ready library — inside Blender, in one pass. Point it at a folder of packs; it detects the vendor, fixes scale/orientation/origin/units, tags every asset, sorts them into tidy category folders, lays them out as a labelled overview grid, and writes a manifest + thumbnail contact sheet.

• Version: 1.0.12 (build file: tagger_v1.0.12.zip)

• Blender: 4.2 or newer

• Licence: GPL-3.0-or-later (see LICENSE)

• Maintainer: Sýdharth Dhandà · Speaking Pictures Entertainment

• Runs 100% local — no cloud, no account, no subscription. Optional AI runs on your machine.

Install

1. Download tagger_v1.0.12.zip.

2. Blender → Edit ▸ Preferences ▸ Add-ons ▸ Install… → pick the zip.

3. Tick Tagger to enable it.

4. Open the sidebar in the 3D Viewport (press N) → Tagger tab.

That's it. Tagger works fully with no extra downloads. The optional local AI is a separate, one-time step — see Optional AI below.

What it does (the workflow)

Tagger's panel is workflow-first: pick one job, follow the numbered steps.

Organise a library — the core job:

Non-destructive and dry-run-first throughout: every "Apply" is preceded by a "Dry-Run" that shows exactly what will happen before anything is written.

Export to Unreal / Unity (one click)

Step 05 · Export to engine turns your tidy library into an engine-ready drop. Pick Unreal, Unity, or Both, choose an output folder, and Tagger writes:

• tagger_library.json — the manifest (every asset + its category, tag and real-world size),

• import_to_unreal.py — a UE Python importer, and/or TaggerLibraryImport.cs — a Unity Editor importer.

Then run the importer in your engine and every asset lands in content folders by category, at the right scale, with collision:

• Unreal: UnrealEditor-Cmd "YourProject.uproject" -ExecutePythonScript="…/import_to_unreal.py" (or Tools ▸ Execute Python Script…). Imports into /Game/TaggerLibrary// — nothing else is touched.

• Unity: put TaggerLibraryImport.cs under Assets/Editor/, then Tools ▸ Tagger ▸ Import Library… and pick the manifest. Imports into Assets/TaggerLibrary//.

Fast mode references your organised files in place; tick Bake clean FBX to re-export each asset through Blender (modifiers applied, scale/orientation baked) for a fully self-contained drop.

Local AI — advanced, arriving in the AI update

Tagger is built to use a local AI (on your own machine, nothing uploaded, no ComfyUI) to identify assets by their thumbnails — a smarter layer on top of the geometry + rules tagging.

This build tags by geometry and folder rules (works fully, no setup). The local-AI layer — a guided one-command install, plus tighter integration so AI tags win over folder guesses — lands in the upcoming AI update. Every AI tag will always be a draft you confirm, never applied blindly.

Then: build a city with it (CityMaker) - Sold Separately

Tagger is step 0 of the Speaking Pictures city pipeline. Once your library has a library_manifest.json, CityMaker (sold separately, and bundled with Tagger) reads it directly to furnish a generated city from your assets:

TAGGER   scan ▸ normalise ▸ tag ▸ sort ▸  library_manifest.json
                                               │
CITYMAKER   ◄── reads the manifest ──  generate layout ▸ POPULATE from your library ▸ roads · rails · surfaces

• Bundled: CityMaker shows a "Step 0 · Tag your library" card that turns green once it finds a valid manifest — then Populate pulls buildings, props and vehicles straight from your tagged library, at correct real-world scale.

• Standalone: the manifest is the whole handshake. Tag in Tagger, then open CityMaker and point it at the same library folder.

You never have to organise your library twice — Tagger does it once, everything downstream reads the result.

Support & links

• Documentation manual: attached with files

• Product documentation page: https://speaking.pictures/tagger.html

• Studio: Speaking Pictures Entertainment

© Speaking Pictures Entertainment. Tagger is licensed under the GNU GPL v3 or later.