# TRELLIS

> Microsoft's TRELLIS image-to-3D (836K+ runs, the most-used 3D model on Replicate) — turns one or more images into a textured GLB 3D asset with turntable render videos.

- **Provider**: replicate
- **Model ID**: firtoz/trellis
- **Category**: 3d_generation
- **Credits**: 77 per request
- **Pricing Type**: fixed

## API Endpoint

Base URL: https://api.core.today/v1

### Create Prediction
POST /predictions

### Get Status
GET /predictions/{job_id}

### Cancel
DELETE /predictions/{job_id}

## Authentication

Header: `X-API-Key: YOUR_API_KEY`

## Input Parameters

- `images` (array, **required**): List of input images to generate 3D asset from
- `slat_guidance_strength` (number, optional): Stage 2: Structured Latent Generation - Guidance Strength (Default: `3`; Range: min: 0, max: 10)
- `seed` (integer, optional): Random seed for generation (Default: `0`)
- `slat_sampling_steps` (integer, optional): Stage 2: Structured Latent Generation - Sampling Steps (Default: `12`; Range: min: 1, max: 50)
- `save_gaussian_ply` (boolean, optional): Save Gaussian point cloud as PLY file (Default: `False`)
- `return_no_background` (boolean, optional): Return the preprocessed images without background (Default: `False`)
- `texture_size` (integer, optional): GLB Extraction - Texture Size (only used if generate_model=True) (Default: `1024`; Range: min: 512, max: 2048)
- `mesh_simplify` (number, optional): GLB Extraction - Mesh Simplification (only used if generate_model=True) (Default: `0.95`; Range: min: 0.9, max: 0.98)
- `randomize_seed` (boolean, optional): Randomize seed (Default: `True`)
- `generate_model` (boolean, optional): Generate 3D model file (GLB) (Default: `False`)
- `generate_normal` (boolean, optional): Generate normal video render (Default: `False`)
- `generate_color` (boolean, optional): Generate color video render (Default: `True`)
- `ss_guidance_strength` (number, optional): Stage 1: Sparse Structure Generation - Guidance Strength (Default: `7.5`; Range: min: 0, max: 10)
- `ss_sampling_steps` (integer, optional): Stage 1: Sparse Structure Generation - Sampling Steps (Default: `12`; Range: min: 1, max: 50)

## Example Request

```json
{
  "model": "firtoz/trellis",
  "input": {
    "images": [
      "https://replicate.delivery/pbxt/MJaYRxQMgIzPsALScNadsZFCXR2h1n97xBzhRinmUQw9aw25/ephemeros_a_dune_sandworm_with_black_background_de398ce7-2276-4634-8f1d-c4ed2423cda4.png"
    ],
    "slat_guidance_strength": 3,
    "seed": 0,
    "slat_sampling_steps": 12,
    "save_gaussian_ply": true,
    "return_no_background": false,
    "texture_size": 2048,
    "mesh_simplify": 0.9,
    "generate_model": true,
    "randomize_seed": true,
    "generate_normal": false,
    "generate_color": true,
    "ss_guidance_strength": 7.5,
    "ss_sampling_steps": 38
  }
}
```

## Response Format

```json
{
  "job_id": "abc123",
  "status": "pending",
  "provider": "replicate",
  "model": "black-forest-labs/flux-schnell",
  "created_at": "2026-01-01T00:00:00Z",
  "result": null,
  "error": null
}
```

Status values: `pending`, `processing`, `completed`, `failed`, `cancelled`

## Usage Flow

1. POST /predictions with model and input → receive job_id
2. Poll GET /predictions/{job_id} until status is `completed` or `failed`
3. Result contains output URL(s) or data

## Output Type

json

## Tags

3d-generation, image-to-3d, glb, trellis, microsoft

## Documentation

https://replicate.com/firtoz/trellis