patina modelsimage generation api

patina/material

Name: PATINA
Author: patina

Material by patina - generate stunning images from text prompts with state-of-the-art AI. Supports multiple aspect ratios, styles, and high-resolution output for creative and commercial use.

Input

Prompt(string)

The text prompt describing the material/texture to generate.

Image URL or upload(string)

Upload

PNG, JPEG, WebP, or GIF · 20 MiB maximum

URL of an input image for image-to-image or inpainting. When provided without mask_url, performs image-to-image; with mask_url, performs inpainting.

This field is required.

Image URL or upload(string)

Upload

PNG, JPEG, WebP, or GIF · 20 MiB maximum

URL of a mask image for inpainting. White regions are regenerated, black regions are preserved. Requires image_url.

Number of Inference Steps(integer)

Number of denoising steps for texture generation. Range: 1 to 8.

Strength(number)

How much to transform the input image. Only used when image_url is provided. Maximum: 1.

Output Format(string)

Output image format for textures and PBR maps. Allowed values: jpeg, png.

Tile Size(integer)

32256

Tile size in latent space (64 = 512px, 128 = 1024px). Range: 32 to 256.

Tiling Mode(string)

Tiling direction: 'both' (omnidirectional), 'horizontal', or 'vertical'. Allowed values: both, horizontal, vertical.

Upscale Factor(integer)

Upscale factor for predicted PBR maps via SeedVR seamless upscaling. 0 = no upscaling, 2 = 2× resolution, 4 = 4× resolution. The base texture image is not upscaled. Allowed values: 0, 2, 4.

Tile Stride(integer)

16128

Tile stride in latent space. Range: 16 to 128.

Maps(array)

Which PBR maps to predict. Deselect all to skip PBR estimation entirely. Defaults to all five.

Aspect Ratio(string)

The aspect ratio of the generated image. Allowed values: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16.

Seed(integer)

Random seed for reproducible generation.

Idle

Example output — click Run to generate your own

Production endpoint

Send your first request

OpenAI-compatible endpoint with unified authentication and usage tracking.

POSThttps://api.sandbase.ai/v1/run

Model IDpatina/material

Input Schema

13 parameters · 2 required · 11 optional

Parameter	Type	Required	Description
`image`	string	Required	URL of an input image for image-to-image or inpainting. When provided without mask_url, performs image-to-image; with mask_url, performs inpainting.
`prompt`	string	Required	The text prompt describing the material/texture to generate.
`maps`	string[]	Optional	Which PBR maps to predict. Deselect all to skip PBR estimation entirely. Defaults to all five. · Default: ["basecolor","normal","roughness","metalness","height"]
`mask`	string	Optional	URL of a mask image for inpainting. White regions are regenerated, black regions are preserved. Requires image_url.
`seed`	integer	Optional	Random seed for reproducible generation.
`strength`	number	Optional	How much to transform the input image. Only used when image_url is provided. · Max: 1 · Default: 0.6
`tile_size`	integer	Optional	Tile size in latent space (64 = 512px, 128 = 1024px). · Min: 32 · Max: 256 · Default: 128
`tile_stride`	integer	Optional	Tile stride in latent space. · Min: 16 · Max: 128 · Default: 64
`tiling_mode`	string	Optional	Tiling direction: 'both' (omnidirectional), 'horizontal', or 'vertical'. · Options: both, horizontal, vertical · Default: "both" bothhorizontalvertical
`aspect_ratio`	string	Optional	The aspect ratio of the generated image. · Options: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16 21:916:93:24:35:41:14:53:42:39:16
`output_format`	string	Optional	Output image format for textures and PBR maps. · Options: jpeg, png · Default: "png" jpegpng
`upscale_factor`	integer	Optional	Upscale factor for predicted PBR maps via SeedVR seamless upscaling. 0 = no upscaling, 2 = 2× resolution, 4 = 4× resolution. The base texture image is not upscaled. · Options: 0, 2, 4 · Default: 0 024
`num_inference_steps`	integer	Optional	Number of denoising steps for texture generation. · Min: 1 · Max: 8 · Default: 8

Output Schema

Field	Type	Description
`id`	string	Unique identifier for the generation task
`status`	string	Task status: pending, running, completed, failed, timeout
`model`	string	Model used for the generation
`outputs`	array	Array of output items
`outputs[].url`	string	URL of the generated artifact
`outputs[].content_type`	string	MIME type (e.g. image/png, video/mp4)
`error`	object \| null	Error details if failed, null on success
`error.type`	string	Machine-readable error type code
`error.message`	string	Human-readable error description

Async Workflow

This model uses asynchronous execution. Submit a request and poll for the result.

Submit — POST to /v1/run, receive an id
Poll — GET /v1/run/{id} until status is completed, failed, or timeout
Retrieve — Read outputs from the completed response

Code Examples

Ready-to-run snippets

# Step 1: Submit
curl -X POST https://api.sandbase.ai/v1/run \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
  "model": "patina/material",
  "maps": [
    "basecolor",
    "normal",
    "roughness",
    "metalness",
    "height"
  ],
  "prompt": "mossy stone wall",
  "strength": 0.6,
  "tile_size": 128,
  "tile_stride": 64,
  "tiling_mode": "both",
  "output_format": "png",
  "upscale_factor": 0,
  "num_inference_steps": 8
}'

# Step 2: Poll result (replace <id>)
curl https://api.sandbase.ai/v1/run/<id> \
  -H "Authorization: Bearer YOUR_API_KEY"

← Back to PATINA

API README

PATINA Material

PATINA Material is the PATINA route for material synthesis, turning a source surface and material brief into a production-ready result while keeping the operation distinct from neighboring endpoints. Material by patina - generate stunning images from text prompts with state-of-the-art AI. Supports multiple aspect ratios, styles, and high-resolution output for creative and commercial use. The workflow is designed for creators who need the model’s specific transformation to remain visible in the request, so the source, intended change, and finished artifact can be reviewed as one coherent creative decision.

In practical use, this route exposes maps, mask, seed, image, prompt, strength, tile_size, tile_stride to shape the exact deliverable. Those controls let a team preserve the important source constraints, state subject behavior or material treatment precisely, choose supported timing or output characteristics, and reproduce successful settings across alternate takes. The result fits an iterative pipeline: establish the core brief, compare controlled variations, then pass the selected asset into editorial, design, localization, visualization, or publishing work.

Highlights

Material-aware synthesis. Create or revise physically plausible surface material from an image and a written material brief.
Production map output. Request the documented material-map set for downstream 3D shading rather than receiving only a flat preview.
Seamless texture control. Use tiling mode, tile size, and stride to prepare repeatable horizontal, vertical, or two-axis textures.
Resolution-linked enhancement. Combine inference strength with the supported 2× or 4× upscale path when the material needs more usable detail.

Pricing

Configuration	Price
Base component	`$0.010000`
Image area	`$0.020000 per megapixel`
Each requested map	`$0.010000 per megapixel`
2× upscale surcharge	`$0.004000 per megapixel per map`
4× upscale surcharge	`$0.016000 per megapixel per map`

When to Use

Scenario	Why it fits
Exact workflow fit	Choose this route when the required deliverable is material synthesis, rather than a related route with different source media.
Directed creative iteration	Use it when subject, motion, material, speech, framing, or finish should be expressed explicitly and compared across controlled variants.
Existing-asset continuity	Use it when supplied images, video, audio, references, or styles must remain the anchor for the generated result.
Repeatable production	Use it when successful inputs need to be saved and rerun across a campaign, asset set, localization pass, or batch.
Pipeline handoff	Use it when the returned artifact will move into editorial, compositing, visualization, review, storage, or publishing.

Prompt Guide

Start with the source or subject, then describe the intended transformation, movement or behavior, camera and composition, and the desired finish. Keep media URLs reachable, use only fields documented for this exact route, and change one major control at a time when comparing results. For source-led tasks, describe what should change as well as what must remain recognizable.

{
  "prompt": "mossy stone wall",
  "image": "https://example.com/reference.jpg",
  "aspect_ratio": "21:9"
}

Technical Specs

Property	Value
Model ID	`patina/material`
Execution mode	`async`
Required inputs	`prompt`, `image`
`maps`	array; default: ['basecolor', 'normal', 'roughness', 'metalness', 'height']
`mask`	string
`seed`	integer
`image`	string
`prompt`	string
`strength`	number; maximum: 1; default: 0.6
`tile_size`	integer; minimum: 32; maximum: 256; default: 128
`tile_stride`	integer; minimum: 16; maximum: 128; default: 64
`tiling_mode`	string; options: both, horizontal, vertical; default: both
`aspect_ratio`	string; options: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16
`output_format`	string; options: jpeg, png; default: png
`upscale_factor`	integer; options: 0, 2, 4; default: 0
`num_inference_steps`	integer; minimum: 1; maximum: 8; default: 8

Related Models

patina/material/extract