3D Generation API Documentation

Single Image Workflow

There are two possible flows when generating from a single image:

Multi-Image Workflow

Similarly, there are two flows for multi-image generation:

Single Image Generation (Python)

import requests
import base64
import time

# API endpoint
BASE_URL = "http://127.0.0.1:7960"

def generate_no_preview(image_base64: str):
    """Generate 3D model from a single base64-encoded image without previews.
    
    Args:
        image_base64: Base64 string of the image (without 'data:image/...' prefix)
    """
    try:
        # Set generation parameters
        params = {
            'image_base64': image_base64,
            'seed': 42,
            'ss_guidance_strength': 7.5,
            'ss_sampling_steps': 30,
            'slat_guidance_strength': 7.5,
            'slat_sampling_steps': 30,
            'mesh_simplify_ratio': 0.95,
            'texture_size': 1024,
            'output_format': 'glb'
        }
        
        # Start generation
        print("Starting generation...")
        response = requests.post(f"{BASE_URL}/generate_no_preview", data=params)
        response.raise_for_status()
        
        # Poll status until complete
        while True:
            status = requests.get(f"{BASE_URL}/status").json()
            print(f"Progress: {status['progress']}%")
            
            if status['status'] == 'COMPLETE':
                break
            elif status['status'] == 'FAILED':
                raise Exception(f"Generation failed: {status['message']}")
            
            time.sleep(1)
        
        # Download the model
        print("Downloading model...")
        response = requests.get(f"{BASE_URL}/download/model")
        response.raise_for_status()
        
        return response.content
        
    except Exception as e:
        print(f"Error: {str(e)}")
        return None

Multi-Image Generation (Python)

def generate_multi_preview(image_base64_list: list[str]):
    """Generate 3D model from multiple base64-encoded images with previews.
    
    Args:
        image_base64_list: List of base64 strings (without 'data:image/...' prefix)
    """
    try:
        # Set generation parameters
        params = {
            'image_list_base64': image_base64_list,
            'seed': 42,
            'ss_guidance_strength': 7.5,
            'ss_sampling_steps': 30,
            'slat_guidance_strength': 7.5,
            'slat_sampling_steps': 30,
            'preview_resolution': 512,
            'preview_frames': 30,
            'preview_fps': 30,
            'mesh_simplify_ratio': 0.95,
            'texture_size': 1024,
            'output_format': 'glb'
        }
        
        # Start generation with preview
        print("Starting generation...")
        response = requests.post(f"{BASE_URL}/generate_multi_preview", data=params)
        response.raise_for_status()
        
        # Poll until preview is ready
        while True:
            status = requests.get(f"{BASE_URL}/status").json()
            print(f"Progress: {status['progress']}%")
            
            if status['status'] == 'PREVIEW_READY':
                break
            elif status['status'] == 'FAILED':
                raise Exception(f"Generation failed: {status['message']}")
            
            time.sleep(1)
        
        # Download preview video
        print("Downloading preview...")
        preview = requests.get(f"{BASE_URL}/download/preview/gaussian")
        preview.raise_for_status()
        
        # Resume generation to get final model
        print("Resuming generation for final model...")
        resume_params = {
            'mesh_simplify_ratio': 0.95,
            'texture_size': 1024
        }
        response = requests.post(f"{BASE_URL}/resume_from_preview", params=resume_params)
        response.raise_for_status()
        
        # Wait for final model
        while True:
            status = requests.get(f"{BASE_URL}/status").json()
            print(f"Progress: {status['progress']}%")
            
            if status['status'] == 'COMPLETE':
                break
            elif status['status'] == 'FAILED':
                raise Exception(f"Final generation failed: {status['message']}")
            
            time.sleep(1)
        
        # Download final model
        print("Downloading final model...")
        model = requests.get(f"{BASE_URL}/download/model")
        model.raise_for_status()
        
        return {
            'preview': preview.content,
            'model': model.content
        }
        
    except Exception as e:
        print(f"Error: {str(e)}")
        return None

Check server status and availability.

{
    "status": "running",
    "message": "Trellis API is operational",
    "busy": boolean
}

Get the status of the current or last generation.

{
    "status": string,
    "progress": number,
    "message": string,
    "busy": boolean
}

Generate a 3D model without previews. Download GLB when complete.

Generate a 3D structure with previews. Download videos when ready.

Generate a 3D model using multiple images without previews.

Generate previews using multiple images.

Continue generation from a preview-ready state to create the final GLB model.

Cancel the current generation process.

Download preview video. Type can be: "gaussian" or "mesh".

Download the final 3D model (GLB format, with texture).