API Reference

Copy Markdown

Open in Claude

Open in ChatGPT

Open in Cursor

---

Copy Markdown

View as Markdown

Search

Asset Search

POST/search/assets

Model Search

POST/search/models

ModelsExpand Collapse

SearchAssetSearchResponse object { id, authorId, collectionIds, 15 more }

id: string

Unique identifier for the asset (e.g., “asset_GTrL3mq4SXWyMxkOHRxlpw”)

authorId: string

User ID of the asset creator/author

collectionIds: array of string

Array of collection IDs this asset belongs to

createdAt: string

Creation timestamp in ISO 8601 format (e.g., “2025-01-16T11:19:41.579Z”)

metadata: object { aspectRatio, baseModelId, guidance, 17 more }

Metadata containing detailed information about the asset’s properties and generation parameters

aspectRatio: optional string

Aspect ratio of the asset (e.g., “1:1”, “16:9”)

baseModelId: optional string

ID of the base model used for generation

guidance: optional number

Guidance scale used for diffusion models to control prompt adherence

height: optional number

Height of the asset in pixels

kind: optional string

Classification of the asset (e.g., “image”, “video”)

modelId: optional string

ID of the specific model used if asset is model-based

modelType: optional string

Type of model used for generation (e.g., “sd-1_5”, “flux.1”)

name: optional string

Custom name given to the asset

negativePrompt: optional string

Negative prompt used during generation to specify what to avoid

negativePromptStrength: optional number

For Flux models: controls the influence of the negative prompt

numInferenceSteps: optional number

Number of inference steps used during generation

parentJobId: optional string

ID of the parent job that created this asset

prompt: optional string

Text prompt used to generate the asset

scheduler: optional string

Scheduler algorithm used during generation

seed: optional string

Random seed used for generation, enables reproducibility

size: optional number

File size of the asset in bytes

text: optional string

Text used to generate the asset: mainly for speech to text

thumbnail: optional object { assetId }

Thumbnail image information for the asset. Ex. Canvas thumbnail

assetId: string

ID of the asset used as thumbnail

type: optional string

Specific type/category of the asset (e.g., “canvas”, “inference-txt”)

width: optional number

Width of the asset in pixels

mimeType: string

MIME type of the asset (e.g., “image/png”, “image/jpeg”)

nsfw: array of string

Array of detected NSFW categories

ownerId: string

Project ID of the asset owner

privacy: string

Privacy setting of the asset (“public”, “private”, or “unlisted”)

tags: array of string

Array of user-assigned tags for categorization

teamId: string

Team ID of the asset owner

updatedAt: string

Last modification timestamp in ISO 8601 format (e.g., “2025-01-16T11:19:41.579Z”)

url: string

Signed URL of the asset

content: optional string

Indexed body for text assets (AssetKind.Text). The full content lives in S3; this field carries up to TEXT_INDEX_MAX_BYTES of UTF-8 for BM25 lookup. Always undefined for non-text assets.

description: optional string

Textual description of the asset. Can be either:

• User-provided description for training images
• Automatically generated caption for images and videos assets
• Automatically generated description for audio assets

score: optional number

Score assigned to the asset based on usage and popularity

thumbnail: optional object { assetId, url }

Thumbnail preview of the asset 3D, video…

assetId: string

ID of the asset used as thumbnail

url: string

Signed URL of the thumbnail

transcription: optional string

Transcription of the asset. Can be either:

• Automatically generated transcription for audio assets with speech

SearchModelSearchResponse object { id, authorId, capabilities, 17 more }

id: string

Unique identifier for the model (e.g., “model_GTrL3mq4SXWyMxkOHRxlpw”)

authorId: string

User ID of the model creator/author

capabilities: array of string

Array of model capabilities

collectionIds: array of string

Array of collection IDs this model belongs to

concepts: array of object { modelId }

Array of concept models associated with this model Each concept contains a reference to its source model

modelId: string

ID of the concept’s source model

createdAt: string

Creation timestamp in ISO 8601 format (e.g., “2025-01-16T11:19:41.579Z”)

exampleAssetIds: array of string

Array of example asset IDs associated with this model

name: string

Display name of the model

ownerId: string

Project ID of the model owner

privacy: string

Privacy setting of the model (“public”, “private”, or “unlisted”)

shortDescription: string

Short description of the model

source: string

Source/origin of the model (e.g., “training”, “import”)

tags: array of string

Array of user-assigned tags for categorization

teamId: string

Team ID of the model owner

trainingImagesNumber: number

Number of images used to train this model

type: string

Type of the model (e.g., “sd-1_5”, “flux.1”)

updatedAt: string

Last modification timestamp in ISO 8601 format (e.g., “2025-01-16T11:19:41.579Z”)

parentModelId: optional string

ID of the parent model this model was derived from

score: optional number

Score assigned to the model based on usage and popularity

thumbnail: optional object { assetId, url }

Thumbnail image information for the model

assetId: string

ID of the asset used as thumbnail

url: string

Signed URL of the thumbnail