Run a prompt once per Asset selected by the specified assets selector, supplying all available context about each asset to the model.
Use the BY_IDS selector to run the prompt for a specific list of Asset.ids, or the BY_CRITERION selector to run it for a search criterion, e.g. all Assets in a specific folder. See AssetSearch for more detail on the available criterion.
At most 500 assets can be selected; larger selections are rejected with a 400.
The completed job's result maps each successfully processed asset ID to the model's output, and each failed asset ID to a failure reason.
The instructions given to the model, run once per selected Asset. All available context about the asset (filename, file type, metadata values, assigned products and variants, and an image of the asset where one is available) is supplied to the model alongside the prompt.
- Productionhttps://api-v2.dash.app/asset-prompt-batch-jobs
- Staginghttps://api-v2.dashstaging.app/asset-prompt-batch-jobs
curl -i -X POST \
https://api-v2.dash.app/asset-prompt-batch-jobs \
-H 'Authorization: Bearer <YOUR_JWT_HERE>' \
-H 'Content-Type: application/json' \
-d '{
"prompt": "Write concise alt text describing the image for screen reader users.",
"selector": {
"type": "BY_CRITERION",
"criterion": {
"type": "FIELD_EQUALS",
"value": "Folder",
"field": {
"type": "FIXED",
"fieldName": "STAGED_BY"
},
"includeDescendants": true,
"boost": null
}
}
}'Success
Unique identifier of the job that has been created
The ID of the account the job is happening in
The ID of the User who created the job
{ "id": "be161977-d44e-4888-af3c-66522e223963", "accountId": "cfb665ca-ce35-4418-b9d5-70ee815db4bd", "creatorId": "google-oauth2|110955770826801837334", "type": "IN_PROGRESS", "status": "PENDING", "progress": { "steps": [ … ] } }