Image to Image

Executes a visual similarity search that takes an input image (provided via coactive_image_id or public_url) and searches within a specified dataset (dataset_id) to find visually similar assets. The input image is encoded into an embedding using the dataset's configured encoder (e.g., CLIP, SigLIP, or other vision-language embedding models), then the embedding is searched against images and video frames in the dataset using vector similarity. Results can be optionally filtered by asset_type (image or video). Returns a ranked list of image and video frame assets ordered by similarity score, including asset metadata, URLs, and optional video information for frame results.

Authentication

AuthorizationBearer

Bearer authentication of the form Bearer <token>, where token is your auth token.

Request

This endpoint expects an object.
dataset_idstringRequiredformat: "uuid"
The unique identifier for the dataset
offsetintegerOptional0-9223372036854776000Defaults to 0
Starting index to return
limitintegerOptional1-1000Defaults to 100
Max number of items to return
metadata_filtersobject or nullOptional
List of metadata filters to apply to the search
coactive_image_idstring or nullOptionalformat: "uuid"
The image to search for represented with a CID
public_urlstring or nullOptional
The image to search for represented with a public http url
upload_idstring or nullOptional
The upload ID to search for
asset_typeenum or nullOptional
The asset type to filter results by. If no asset type is selected the endpoint will search over images and videos
Allowed values:

Response

Successful Response
datalist of objects
The paginated results

Errors

422
Unprocessable Entity Error