3D Model Pre-flight: GLB Checker & Optimizer
Pricing
from $20.00 / 1,000 model checkeds
3D Model Pre-flight: GLB Checker & Optimizer
Check whether an AI-generated 3D model is print-ready or game-ready, then download an optimized version. Detects non-watertight meshes, missing tangents, over-budget geometry and textures.
Pricing
from $20.00 / 1,000 model checkeds
Rating
0.0
(0)
Developer
dada gou
Maintained by CommunityActor stats
0
Bookmarked
2
Total users
1
Monthly active users
6 days ago
Last modified
Categories
Share
Check whether an AI-generated 3D model is actually usable — before it breaks your pipeline or wastes a print.
AI generators (Meshy, Tripo, Hunyuan3D) produce models that look fine but are often not usable: meshes with thousands of open edges that no 3D printer can accept, missing tangent data, geometry 30x over a game budget. Opening each file in Blender to find out does not scale.
This Actor checks a .glb file for you and — if you want — hands back an optimized copy.
What you get
For every model:
- A plain-language verdict —
game_ready,needs_fixes, ornot_usable - A list of concrete problems — e.g. "Mesh has 68,746 open boundary edges — not watertight. It cannot be sent to a 3D printer as-is."
- An optimized
.glb(optional) — real-world results: 88–90% smaller, still valid per the official glTF validator
Input
| Field | What to put |
|---|---|
| Model URL | A direct link to a .glb file, or to a .zip containing up to 50 .glb files. Leave the default bundled:sample to try the Actor on the small sample model that ships with it |
| Target platform budget | mobile_prop / pc_prop / character / hero_asset — or report_only for statistics without pass/fail |
| Also produce an optimized file | On by default. Turn off if you only want the report |
| Texture format | keep (safest — works in any engine) or webp (smaller, needs engine support) |
| Max models from a zip | Cost guard, default 10 |
What it detects
| Check | Why it matters |
|---|---|
| Not watertight | Open edges mean a 3D printer cannot build it |
| Non-manifold edges | Slicers and engines behave unpredictably |
| Degenerate triangles | Shading artifacts |
| Missing tangents | Normal maps render differently (or slower) in every engine |
| Missing normals / UVs | Lighting or textures will not display correctly |
| Over triangle budget | Too heavy for the target platform |
| Over texture budget | Too much VRAM |
| Uncompressed textures | File size can be cut further |
| glTF specification errors | Some engines will refuse to load the file |
How to read the result
Every model produces one row in the run's dataset:
verdict— the one-word answerissues[]— each problem, in buyer language, with a severity (error/warn/info)geometry,topology,materials— the raw numbers behind the verdictoptimized— download key, size before/after, and the verification result
A batch summary is stored under the OUTPUT key.
Notes
- Optimized files are written to this run's key-value store; download them from the run's Output tab.
- Every optimized file is re-checked with the official Khronos glTF validator before delivery. If a file fails verification, it is not delivered.
- Draco compression is lossy by a very small amount (typically under 0.02% of triangles). Use
optimize: falseif you need bit-exact geometry.
Pricing
Pay-per-event:
| Event | When |
|---|---|
| Run start | once per run |
| Model checked | per model analyzed |
| Model optimized | per optimized file delivered |
| Size surcharge | per 10 MB of input above 10 MB |
Billing events are triggered as the work actually happens: the run start is charged immediately, each model is charged once it has been checked, and an optimized file is charged only if it was delivered.
