Model Compression

The iCraft Online Model Compression Tool allows you to compress GLB/GLTF 3D models with one click, reducing file size by up to 95% while preserving visual quality as much as possible.

This tool runs directly in the browser, requires no software installation, and provides all advanced features for free.

Supported Formats and Use Cases

• Supported Formats

  • GLB

  • GLTF

  Typical Use Cases

  • Preparing models for Web or Player publishing

  • Optimizing models for mobile devices

  • Embedding models in web pages

  • Reducing loading time

  • Saving bandwidth and storage

  • Improving the smoothness of project demonstrations

  Large Model Recommendation

  • Browser-based processing is limited by device memory.

  • In most cases, models below 100MB can be processed smoothly.

  • For extremely large models, it is recommended to simplify the geometry first using external tools before compression.

---

Key Features

• No Installation, No Upload Required All compression is performed locally in the browser, ensuring privacy and data security.

• Completely Free All advanced compression options are available without registration or payment.

• High Compatibility Compressed files remain fully compliant with GLB/GLTF standards, ensuring compatibility with platforms such as Three.js、Unity、Unreal.

• Advanced Compression Options Supports adjustable parameters including texture compression, geometry compression, and attribute quantization.

---

Quick Start Guide

1

Upload the Model

Click or drag a GLB/GLTF file into the upload area. The uploaded file will appear in the file list.

2

Configure Compression Parameters

• Texture Format Available options: WebP、JPEG、PNG、AVIF， We recommend to use WebP, can significantly reduces texture size while maintaining compatibility with most browsers.

• Texture Size Available options: 512x512、1024x1024、2048x2048、4096x4096 Default size: 512x512 This option downscales texture size.

• Texture Quality Adjust the texture compression quality via the slider.The lower the value, the less the texture quality is compressed. Default value: 80 Adjustable range: 1~100

3

(Optional) Advanced Options

Adjust compression parameters for geometry and attributes (e.g., encoding/decoding speed, position, normal, texture coordinate, color quantization, etc.). These settings are used to further reduce vertex count and attribute precision, thereby compressing file size.

4

Start Compression

Click the Compress button. Processing time depends on model complexity: small models can be completed in seconds, while complex models may take tens of seconds to several minutes.

5

Download

Download the compressed GLB/GLTF file once compression is finished.

6

Verify Compatibility

Load the model on target platforms (e.g., iCraft Player, Three.js, Unity, Blender) and quickly check that model display, textures, and animations are working correctly. Compressed models are generally compatible with mainstream engines and 3D design software.

---

Practical Tips

1. Always back up the original file: Keep an uncompressed original for rollback purposes.

2. Compress in stages: Start with mild compression (e.g., reduce texture quality by one level + enable deduplication), and only apply aggressive settings after verifying the result.

3. Prioritize texture optimization: Textures account for most of the model size. Use texture format (WebP), resolution reduction and quality compression for the greatest size reduction.

4. Validate animations and rigs on the target platform: If the model contains skeletal animations or custom materials, always test them in the target engine.

5. Use advanced options with caution: Position/normal quantization and vertex welding affect model precision. They are suitable for presentation/demo scenes, but should be used carefully for precision engineering applications.

6. Split large scenes for better performance: Break oversized scenes into sub-models, compress them separately, then recombine.