Skip to main content
POST
cURL

Usage notes

  • Send either schema_id or columns, not both. Lasso uses the default Product Schema when you omit both.
  • A request may contain up to 50 products. Send 1 to 5 products for a synchronous response. For 6 to 50 products, you must provide webhook_url because there is no public endpoint for retrieving an Enrich job result.
  • thinking costs 4 credits per product for hard, 2 for medium, and 1 for low.
  • With web_search: false, Lasso uses model knowledge and does not return citations in basis.
  • Use context for instructions that apply to every item in the request.
  • model is a deprecated compatibility field. The service ignores it; thinking selects the service-controlled model tier.
  • In synchronous mode, insufficient credits currently produce HTTP 200 with error: "Insufficient credits", no enriched items, and credits_used: 0.
Do not send more than 5 items without webhook_url. The current handler accepts the request and returns 202, but it exposes no result-retrieval route for that job ID.

Examples

Example response

Authorizations

Authorization
string
header
required

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

Body

application/json
items
object[]
required

Send at most 5 items for a synchronous response. Requests with 6 to 50 items require webhook_url to make the completed result retrievable.

Required array length: 1 - 50 elements
schema_id
string

ID of an existing schema. Mutually exclusive with columns.

columns
object[]

Inline column definitions. Mutually exclusive with schema_id.

context
string

Additional context for the AI enrichment

model
string
deprecated

Legacy compatibility field. The service ignores the requested value; thinking selects the service-controlled model tier.

use_glossary
boolean
default:false

Whether to use web search for enrichment. When false, the AI fills fields from its own knowledge and no basis/citations are returned.

thinking
enum<string>
default:medium

Controls the depth of AI reasoning. hard — most capable model, best for complex or ambiguous products. medium — balanced speed and quality (default). low — fastest, suitable for straightforward lookups.

Available options:
hard,
medium,
low
webhook_url
string<uri>

Delivers results asynchronously and returns 202. Required for 6 to 50 items because there is no public Enrich job-result endpoint.

Response

Enrichment results for 1 to 5 items in synchronous mode. A request-wide failure such as insufficient credits can currently appear as an error field in this HTTP 200 response.

id
string
status
enum<string>
Available options:
completed
items
object[]
credits_used
number
error
string

Present for a request-wide worker failure such as insufficient credits. The current synchronous handler can still return this payload with HTTP 200.