- Home
- Skills
- Frontend & Web Apps
- Glb Web Optimizer
Works with the AI Agents you already use
Glb Web Optimizer
Audit and optimize GLB/glTF models for the web using Draco, Meshopt, and WebP texture compression.
Free
Glb Web Optimizer
Example session with this skill installed
My hero_banner.glb is 15MB and making the site slow. Can you audit it and then optimize it for a landing page? Use Draco compression and resize textures.
- Read your context and instructions
- Compiled the glb web optimizer
I have audited the model and performed the optimization. The file size has been reduced by 78% while staying within the 'hero' budget.
Optimization Results
| Metric | Original (hero_banner.glb) | Optimized (hero_banner.optimized.glb) | Status |
|---|---|---|---|
| File Size | 15.42 MB | 3.24 MB | Close to Budget |
| Triangles | 182,400 | 182,400 | Over Budget |
| Draw Calls | 64 | 64 | Over Budget |
| Max Texture | 4096 px | 2048 px | Within Budget |
| Texture Format | PNG/JPEG | WebP | Optimized |
| GPU Memory (Est) | 112 MB | 28 MB | Optimized |
The optimization converted all textures to WebP and capped them at 2048px. I applied Draco geometry compression which significantly reduced the transfer size. Note that the triangle count is still above the 100k hero budget; if you need further reduction, we can run a simplification pass at the cost of some geometric detail.
Next steps
- Inspect the model: Open
hero_banner.optimized.glbin a viewer to ensure texture quality is acceptable.
Update your loader: Ensure your Three.js GLTFLoader is configured with a DRACOLoader instance to decode the mesh.
3.
Further simplify: If the page is still sluggish on mobile, I can run node scripts/optimize.mjs --simplify 0.5 to halve the triangle count.
Connects securely to your tools. The creator never sees your data.
About this skill
Large 3D models break web performance, leading to slow page loads and high memory usage. This skill audits glTF and GLB models against web-performance budgets and runs automated optimization pipelines to shrink them. It provides objective data on triangle counts, draw calls, and GPU memory usage before and after processing.
What it does
- Perform audits by checking models against hero or configurator performance budgets for size and complexity.
- Compress geometry using Draco or Meshopt to significantly reduce the transfer size of meshes.
- Optimize textures by converting to WebP, resizing dimensions, and pruning redundant assets.
- Reduce complexity via mesh simplification and joining to lower draw calls on static models.
- Estimate GPU memory to predict the actual impact on the user's hardware beyond just file size.
- Generate reports showing a side-by-side comparison of metrics between the original and optimized files.
How it works
- Analyze the source by running the audit script on a
.glbor.gltffile to identify performance bottlenecks. - Select a profile such as hero or configurator to establish the target budget for the optimization.
- Run optimization using specific flags for animation support, texture quality, or mesh simplification.
- Verify the output by auditing the new file to ensure it meets the target performance criteria.
Frameworks & tools
The skill uses Node.js scripts powered by gltf-transform and the sharp image processing library. It provides implementation snippets for Three.js and React Three Fiber.
Why this beats prompting it yourself
Generic AI prompting cannot inspect the binary structure of a GLB file or calculate actual GPU memory footprints. This skill uses local scripts to perform real audits and deterministic optimizations, ensuring the resulting files are valid and performant.
Use cases
- Shrinking landing page assets to improve Core Web Vitals and load speeds.
- Preparing product configurators that require high-resolution textures without crashing mobile browsers.
- Optimizing animated avatars for social or gaming applications using Meshopt compression.
- Fixing draw call bottlenecks in complex scenes by joining static meshes.
Known limitations
Does not fix UV maps, bake lighting, or generate KTX2 textures. Requires Node.js 18 or newer to run the bundled scripts.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Frontend & Web Apps.
Requires Node.js 18+. Does not support KTX2 generation, UV editing, or light baking. Optimization requires a local npm install of dependencies.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account
- Accounts
GitHub
Opencollective
- Software
Node.js
Permissions
Allowed Hosts
File Scopes