Trellis API

Trellis
Multi

Endpoint: POST https://fal.run/fal-ai/trellis Endpoint ID: fal-ai/trellis

Try it in the Playground

Run this model interactively with your own prompts.

Quick Start

import fal_client

def on_queue_update(update):
    if isinstance(update, fal_client.InProgress):
        for log in update.logs:
           print(log["message"])

result = fal_client.subscribe(
    "fal-ai/trellis",
    arguments={
        "image_url": "https://storage.googleapis.com/falserverless/web-examples/rodin3d/warriorwoman.png"
    },
    with_logs=True,
    on_queue_update=on_queue_update,
)
print(result)

Input Schema

image_url

string

required

URL of the input image to convert to 3D

seed

integer

Random seed for reproducibility

ss_guidance_strength

float

default:"7.5"

Guidance strength for sparse structure generation Default value: 7.5Range: 0 to 10

ss_sampling_steps

integer

default:"12"

Sampling steps for sparse structure generation Default value: 12Range: 1 to 50

slat_guidance_strength

float

default:"3"

Guidance strength for structured latent generation Default value: 3Range: 0 to 10

slat_sampling_steps

integer

default:"12"

Sampling steps for structured latent generation Default value: 12Range: 1 to 50

mesh_simplify

float

default:"0.95"

Mesh simplification factor Default value: 0.95Range: 0.9 to 0.98

texture_size

TextureSizeEnum

default:"1024"

Texture resolution Default value: "1024"Possible values: 512, 1024, 2048

Output Schema

model_mesh

File

required

Generated 3D mesh file

timings

Timings

required

Processing timings

Input Example

{
  "image_url": "https://storage.googleapis.com/falserverless/web-examples/rodin3d/warriorwoman.png",
  "ss_guidance_strength": 7.5,
  "ss_sampling_steps": 12,
  "slat_guidance_strength": 3,
  "slat_sampling_steps": 12,
  "mesh_simplify": 0.95,
  "texture_size": 1024
}

Output Example

{
  "model_mesh": {
    "url": "",
    "content_type": "image/png",
    "file_name": "z9RV14K95DvU.png",
    "file_size": 4404019
  }
}

Endpoint: POST https://fal.run/fal-ai/trellis/multi Endpoint ID: fal-ai/trellis/multi

Try it in the Playground

Run this model interactively with your own prompts.

Quick Start

import fal_client

def on_queue_update(update):
    if isinstance(update, fal_client.InProgress):
        for log in update.logs:
           print(log["message"])

result = fal_client.subscribe(
    "fal-ai/trellis/multi",
    arguments={
        "image_urls": [
            "https://storage.googleapis.com/falserverless/model_tests/video_models/front.png",
            "https://storage.googleapis.com/falserverless/model_tests/video_models/back.png",
            "https://storage.googleapis.com/falserverless/model_tests/video_models/left.png"
        ]
    },
    with_logs=True,
    on_queue_update=on_queue_update,
)
print(result)

Input Schema

image_urls

list<string>

required

List of URLs of input images to convert to 3D

seed

integer

Random seed for reproducibility

ss_guidance_strength

float

default:"7.5"

Guidance strength for sparse structure generation Default value: 7.5Range: 0 to 10

ss_sampling_steps

integer

default:"12"

Sampling steps for sparse structure generation Default value: 12Range: 1 to 50

slat_guidance_strength

float

default:"3"

Guidance strength for structured latent generation Default value: 3Range: 0 to 10

slat_sampling_steps

integer

default:"12"

Sampling steps for structured latent generation Default value: 12Range: 1 to 50

mesh_simplify

float

default:"0.95"

Mesh simplification factor Default value: 0.95Range: 0.9 to 0.98

texture_size

TextureSizeEnum

default:"1024"

Texture resolution Default value: "1024"Possible values: 512, 1024, 2048

multiimage_algo

MultiimageAlgoEnum

default:"stochastic"

Algorithm for multi-image generation Default value: "stochastic"Possible values: stochastic, multidiffusion

Output Schema

model_mesh

File

required

Generated 3D mesh file

timings

Timings

required

Processing timings

Input Example

{
  "image_urls": [
    "https://storage.googleapis.com/falserverless/model_tests/video_models/front.png",
    "https://storage.googleapis.com/falserverless/model_tests/video_models/back.png",
    "https://storage.googleapis.com/falserverless/model_tests/video_models/left.png"
  ],
  "ss_guidance_strength": 7.5,
  "ss_sampling_steps": 12,
  "slat_guidance_strength": 3,
  "slat_sampling_steps": 12,
  "mesh_simplify": 0.95,
  "texture_size": 1024,
  "multiimage_algo": "stochastic"
}

Output Example

{
  "model_mesh": {
    "url": "",
    "content_type": "image/png",
    "file_name": "z9RV14K95DvU.png",
    "file_size": 4404019
  }
}

Trellis - Advanced AI Model for 3D Generation

Transform your creative workflows with Trellis, a powerful AI model designed for reliable, high-quality 3D asset generation from images. Built for developers who need production-ready 3D models with minimal complexity.

Overview

Trellis provides a streamlined API for generating 3D models from images, utilizing a native 3D generative model based on Structured LATents (SLAT) representation. Whether you’re building creative tools, content generation systems, or enhancing existing applications, Trellis delivers consistent, high-quality 3D assets with simple integration.

Getting Started

Setting up Trellis takes just a few minutes. Choose your preferred SDK: For JavaScript/TypeScript projects:

npm install --save @fal-ai/client

For Python applications:

pip install fal-client

Configure your authentication: JavaScript:

import { fal } from "@fal-ai/client";

fal.config({
  credentials: "YOUR_FAL_KEY_HERE"
});

Python:

import fal_client
import os

os.environ["FAL_KEY"] = "YOUR_FAL_KEY_HERE"

Basic Usage

Generate your first 3D model with just a few lines of code: JavaScript:

const result = await fal.subscribe("fal-ai/trellis", {
  input: {
    image_url: "https://example.com/your-image.jpg"
  }
});

Python:

result = fal_client.subscribe("fal-ai/trellis", {
    "image_url": "https://example.com/your-image.jpg"
})

Advanced Features

Trellis supports advanced configuration options to fine-tune your results: 3D Output Formats

Generate multiple 3D representations (Radiance Fields, 3D Gaussians, meshes)
Export to GLB format for universal compatibility
Maintain high detail in both geometry and texture

Quality Control

Intricate shape and texture detail preservation
Flexible output format selection
Support for local 3D editing capabilities

API Parameters

Input Parameters:

image_url (required): URL of the input image
Additional configuration options available for advanced use cases

Output:

3D model in multiple formats
GLB file export capability
High-quality mesh and texture data

Best Practices

Maximize your success with Trellis by following these guidelines: Error Handling

try {
  const result = await fal.subscribe("fal-ai/trellis", {
    input: { image_url: "your-image-url" }
  });
} catch (error) {
  console.error("Generation failed:", error.message);
}

Image Requirements

Use clear, well-lit images with distinct subjects
Supported formats: JPG, JPEG, PNG, WebP, GIF, AVIF
Best results with images showing full object visibility

Technical Specifications

Model Architecture:

Based on Structured LATents (SLAT) representation
Powered by Rectified Flow Transformers
Up to 2 billion parameters
Trained on 500K diverse 3D objects

Performance Metrics:

Fast 3D generation from single images
High-quality output with intricate details
Multiple output format support

File Upload Support

Upload local images using the file storage API:

import { fal } from "@fal-ai/client";

const file = new File([imageData], "image.jpg", { type: "image/jpeg" });
const url = await fal.storage.upload(file);

const result = await fal.subscribe("fal-ai/trellis", {
  input: {
    image_url: url
  }
});

Queue-Based Processing

For asynchronous workflows:

// Submit request
const { request_id } = await fal.queue.submit("fal-ai/trellis", {
  input: {
    image_url: "your-image-url"
  },
  webhookUrl: "https://optional.webhook.url/for/results"
});

// Check status
const status = await fal.queue.status("fal-ai/trellis", {
  requestId: request_id,
  logs: true
});

// Get result
const result = await fal.queue.result("fal-ai/trellis", {
  requestId: request_id
});

Integration Support

Our documentation includes complete reference implementations for:

React/Next.js applications
Python backend services
Node.js environments
REST API clients

Use Cases

3D Asset Creation

Game development assets
Product visualization
Digital art and design
AR/VR content creation

Professional Applications

E-commerce product models
Architecture visualization
Educational content
Creative prototyping

Pricing and Usage

Cost: $0.02 per unit
Transparent, usage-based pricing
No subscription necessary
No hidden fees or minimum commitments

View detailed pricing or contact sales for enterprise solutions.

Support and Resources

Access comprehensive support through multiple channels: Documentation

API Reference
General fal.ai Documentation
Code examples and tutorials

Community Support

Active developer community
Regular updates and improvements
Direct technical support

Get started with Trellis today and experience the next generation of AI-powered 3D asset generation. Sign up for an API key at fal.ai.

Trellis — 3D

Limitations

ss_guidance_strength range: 0 to 10
ss_sampling_steps range: 1 to 50
slat_guidance_strength range: 0 to 10
slat_sampling_steps range: 1 to 50
mesh_simplify range: 0.9 to 0.98
texture_size restricted to: 512, 1024, 2048
multiimage_algo restricted to: stochastic, multidiffusion

Video Generation

Image Generation

Audio

Vision

3D

Trellis API

Try it in the Playground

Quick Start

Input Schema

Output Schema

Input Example

Output Example

Try it in the Playground

Quick Start

Input Schema

Output Schema

Input Example

Output Example

Trellis - Advanced AI Model for 3D Generation

Overview

Getting Started

Basic Usage

Advanced Features

API Parameters

Best Practices

Technical Specifications

File Upload Support

Queue-Based Processing

Integration Support

Use Cases

Pricing and Usage

Support and Resources

Limitations

Video Generation

Image Generation

Audio

Vision

3D

Try it in the Playground

​Quick Start

​Input Schema

​Output Schema

​Input Example

​Output Example

Try it in the Playground

​Quick Start

​Input Schema

​Output Schema

​Input Example

​Output Example

​Trellis - Advanced AI Model for 3D Generation

​Overview

​Getting Started

​Basic Usage

​Advanced Features

​API Parameters

​Best Practices

​Technical Specifications

​File Upload Support

​Queue-Based Processing

​Integration Support

​Use Cases

​Pricing and Usage

​Support and Resources

​Related

​Limitations

Quick Start

Input Schema

Output Schema

Input Example

Output Example

Quick Start

Input Schema

Output Schema

Input Example

Output Example

Trellis - Advanced AI Model for 3D Generation

Overview

Getting Started

Basic Usage

Advanced Features

API Parameters

Best Practices

Technical Specifications

File Upload Support

Queue-Based Processing

Integration Support

Use Cases

Pricing and Usage

Support and Resources

Related

Limitations