Kolbo.AIKolbo.AI Docs

3D Tools

Generate 3D models from single images, multiple images, or text descriptions

3D Tools

Generate 3D models and assets using Kolbo.AI's 3D generation tools with three different input modes.

Overview

Kolbo.AI's 3D Tools provide three modes for creating 3D models:

  • Single Image to 3D: Generate 3D from one image (max 1 image)
  • Multi-Image to 3D: Create 3D from multiple angles (max 4 images)
  • Text to 3D: Generate 3D models from text descriptions

Generation Modes

Single Image to 3D

Generate a 3D model from a single image:

  • Upload 1 image
  • AI infers 3D structure from the single view
  • Generates complete 3D model
  • Fast workflow

Best for:

  • Product shots with clear views
  • Objects with simple geometry
  • Quick 3D mockups
  • Single-angle references

Multi-Image to 3D

Create more accurate 3D from multiple angles:

  • Upload up to 4 images
  • Different angles and perspectives of the same object
  • AI reconstructs 3D from multiple views
  • More accurate and detailed results

Best for:

  • Detailed products
  • Complex objects
  • Photorealistic models
  • Professional 3D assets

Tips:

  • Capture object from different angles
  • Consistent lighting across images
  • Include front, side, and back views
  • 4 images recommended for best results

Text to 3D

Generate 3D models from text descriptions:

  • Describe the object in text
  • AI generates 3D model from description
  • No images required
  • Creative freedom

Best for:

  • Concept visualization
  • Imagined objects
  • Quick ideation
  • Creative projects without reference images

Supported File Formats

Image Upload:

  • JPEG/JPG
  • PNG
  • WebP
  • HEIC/HEIF

Maximum file size: 20MB per image

Advanced Settings

T-Pose / A-Pose

Generate the model in A-pose or T-pose:

  • Ideal for character models
  • Designed for rigging and animation
  • Makes character animation easier
  • Toggle on/off

Target Polycount

Control model detail level:

  • Adjustable polygon count
  • Default: 30,000 polygons
  • Higher = more detail, larger file
  • Lower = faster, smaller file

PBR (Physically Based Rendering)

Generate realistic material maps:

  • Metallic maps
  • Roughness maps
  • Normal maps
  • Realistic material properties
  • Must be OFF for sculpture art style

Prompt Expansion

AI enhancement of your text prompt:

  • AI expands your description
  • Adds detail for better results
  • Improves generation quality
  • Toggle on/off

Topology

Mesh topology type:

  • Triangle: Standard topology
  • Quad: Quad-based topology (better for animation/subdivision)

Symmetry

Model symmetry options:

  • Off: No symmetry enforced
  • Auto: AI decides if symmetry is needed
  • On: Force symmetrical model

Art Style

Visual style of the model:

  • Realistic: Photorealistic materials and textures
  • Sculpture: Artistic sculpture style (PBR must be OFF)

Texture Prompt

Optional prompt for textures:

  • Describe desired textures
  • Separate from main prompt
  • Refine material appearance
  • Optional field

Generation Workflow

Single Image Workflow

  1. Select "Single Image to 3D" mode
  2. Upload your image (or select from Media Library)
  3. Enter optional prompt for guidance
  4. Configure settings (T-pose, polycount, PBR, etc.)
  5. Choose AI model
  6. Click Generate
  7. Wait for processing
  8. View, download, or use the 3D model

Multi-Image Workflow

  1. Select "Multi-Image to 3D" mode
  2. Upload up to 4 images from different angles
  3. Enter optional prompt
  4. Configure settings
  5. Choose AI model
  6. Click Generate
  7. Processing time varies with image count
  8. View and export result

Text to 3D Workflow

  1. Select "Text to 3D" mode
  2. Write descriptive prompt
  3. Optionally add texture prompt
  4. Configure settings
  5. Choose AI model
  6. Click Generate
  7. AI creates model from description
  8. View and download result

AI Models

Multiple AI models available for 3D generation:

  • Different models for each mode
  • Single image models
  • Multi-image models
  • Text-to-3D models
  • Various quality/speed tradeoffs
  • Credit costs vary by model

Viewing & Exporting

3D Viewer

Built-in 3D viewer:

  • Rotate and inspect model
  • Zoom in/out
  • Check from all angles
  • Verify quality before download

Download Options

Export your 3D models:

  • Standard 3D file formats
  • Ready for use in 3D software
  • Compatible with major tools
  • High-quality exports

Generation History

Access your generated models:

  • View all past generations
  • Organized by project
  • Download previous models
  • Manage your 3D library

Use Cases

Product Visualization

  • E-commerce 3D previews
  • Product configuration
  • AR product try-on
  • Interactive catalogs

Game Assets

  • Character models
  • Props and objects
  • Environment assets
  • Game-ready meshes

Architecture & Design

  • Furniture models
  • Decor objects
  • Product prototypes
  • Concept visualization

Animation & Film

  • Character bases for rigging
  • Asset creation
  • Concept models
  • Pre-visualization

AR/VR

  • AR experiences
  • VR environments
  • Interactive 3D content
  • Immersive applications

Best Practices

Image Preparation

For Best Results:

  • High-resolution images
  • Clear object focus
  • Good, even lighting
  • Clean background
  • Multiple angles (multi-image mode)

Avoid:

  • Blurry images
  • Poor lighting
  • Complex backgrounds
  • Extreme angles

Prompt Writing

Effective Text-to-3D Prompts:

  • Be specific and detailed
  • Describe shape and proportions
  • Mention materials and textures
  • Include style references
  • Add detail keywords

Example prompts:

  • "A red ceramic coffee mug with a smooth glossy finish and a curved handle"
  • "Modern minimalist wooden chair with straight legs and a cushioned seat"
  • "Fantasy sword with ornate gold handle and engraved steel blade"

Settings Selection

Choose appropriate settings:

  • T-pose for characters that will be animated
  • Higher polycount for detailed models
  • PBR on for realistic materials
  • PBR off for sculpture art style
  • Quad topology for animation workflows

Tips for Success

  1. Start Simple: Begin with single image mode to learn
  2. Quality Input: Use high-quality, clear images
  3. Multiple Angles: Use multi-image for better results
  4. Clear Descriptions: Detailed prompts for text-to-3D
  5. Test Settings: Experiment with different configurations
  6. Check Preview: Review in 3D viewer before downloading
  7. Iterate: Generate variations if needed
  8. Optimize: Adjust polycount for your use case

Limitations

Image Quality

  • Better input images = better results
  • Low-quality images may produce poor models
  • Lighting affects reconstruction quality

Complex Objects

  • Very complex objects may be simplified
  • Transparent/reflective materials challenging
  • Best results with solid, well-defined objects

Processing Time

  • Single image: Fastest
  • Multi-image: Longer (more images = more time)
  • Text-to-3D: Varies by complexity
  • High polycount: Takes longer

File Sizes

  • Higher polycount = larger files
  • PBR maps add to file size
  • Consider target platform requirements

Troubleshooting

Generation Failed:

  • Check image quality and format
  • Verify file size under 20MB
  • Try different AI model
  • Simplify prompt (text-to-3D)
  • Contact support if issue persists

Poor Results:

  • Use higher quality images
  • Try multi-image mode for better accuracy
  • Adjust settings (polycount, PBR)
  • Write more detailed prompts
  • Generate multiple variations

Model Issues:

  • Check in 3D viewer first
  • Verify all angles look correct
  • Try different topology setting
  • Adjust polycount if needed
  • Use PBR appropriately

Credit Costs

3D generation consumes credits based on:

  • Selected AI model
  • Generation mode
  • Settings complexity
  • Credit cost displayed before generation

Check your subscription plan for specific costs per generation.