Classification Endpoints
The API offers 3 classification levels (workflows) with different speed/accuracy tradeoffs. All endpoints use POST method.
/classify/queryBasic
Direct search without AI enrichment. Ideal for fast vector search or prototyping.
POST
/api/v1/classify/queryRecherche vectorielle simple
Body Parameters
| Name | Type | Description |
|---|---|---|
labels required | array | List of products to classify |
labels[].label required | string | Product description |
labels[].country optional | string | ISO country code |
base required | string | Collection to use |
top_k optional | integerDefault: 3 | Number of results per product |
Example Request
Response
200
Example Response/classify/enrichRecommandé
Classification with intelligent AI label enrichment before search. Significantly improves accuracy by adding technical context and customs vocabulary.
POST
/api/v1/classify/enrichClassification avec enrichissement IA
Body Parameters
| Name | Type | Description |
|---|---|---|
labels required | array | List of products to classify |
labels[].label required | string | Product description |
labels[].country optional | string | ISO country code |
base required | string | Collection to use |
top_k optional | integerDefault: 3 | Number of results per product |
Example Request
Response
200
Example Response/classify/validateExpert
High-end classification with double validation. An expert AI analyzes candidates and selects the best code with detailed justification. Recommended for critical decisions.
POST
/api/v1/classify/validateValidation experte avec justification
Body Parameters
| Name | Type | Description |
|---|---|---|
labels required | array | List of products to classify |
labels[].label required | string | Product description |
labels[].country optional | string | ISO country code |
base required | string | Collection to use |
top_k optional | integerDefault: 3 | Number of results per product |
Example Request
Response
200
Example Response