Current verdict

Waiting for package

Import models and textures first, then follow the check results into the right tool.

Start QA

Pipeline

6 STEPS · LEFT TO RIGHT

Model check

PBR texture pairing

Risk check

SIZE · GEOMETRY · TEXTURE

    glTF spec validation

    KHRONOS VALIDATOR · ON-DEMAND

    Validates GLB / GLTF against the spec using the Khronos official glTF Validator (the validator bundle, ~610 KB, is lazy-loaded on first use).

    Detailed inventory

    MODELS · TEXTURES · HDRI · REFS

    Models

    Textures

    HDRI

    Refs / deps

    FAQ

    FAQ · 6 QUESTIONS
    How long does the 3D Asset Package QA workflow take?

    Small packs under 5MB report in 30 seconds; about 100MB packs take 2-3 minutes; over 200MB the browser may hang, in which case pre-optimize with desktop CLIs (gltf-transform) first.

    Are the files I drop in uploaded to a server?

    No. The entire workflow (Three.js parsing, @gltf-transform optimization, Khronos validator, Canvas pixel analysis) runs locally in your browser. Files never leave your machine and the workflow is NDA-safe.

    What is the Khronos glTF Validator and why is it lazy-loaded?

    The official spec validator maintained by Khronos (github.com/KhronosGroup/glTF-Validator). Sketchfab, Google AR and others use it to gate GLB uploads. We bundle it as a 610KB separate file, downloaded only when you click Run spec validation, so first-time visitors do not pay the cost.

    What does the stale label on a report mean?

    Each report section is tagged with the asset-package version it was generated against. After you replace a model and re-check, old sections auto-flip to a yellow stale badge so you never hand off an out-of-date report.

    Can I keep multiple clients packages side by side?

    Yes. Click the package badge in the cmdbar to open the dropdown, up to 20 packages with switch / rename / clone / delete. Compare two via the Asset Package Diff page (diff.html).

    What is the difference between JSON / Markdown / HTML report formats?

    JSON is for machines (CI / automation); Markdown is great for Jira / Notion / GitHub issues; HTML is a standalone email-ready report page. Same data, three formats.