RS Infer
Server Endpoints

/v1/score

Score the `text_1 × text_2` cross-product (vLLM /v1/score). Every pair between `text_1` and `text_2` is scored with the reranker. The response `data` array has `|text_1| × |text_2|` items in pair order.

POST
/v1/score

Score the text_1 × text_2 cross-product (vLLM /v1/score). Every pair between text_1 and text_2 is scored with the reranker. The response data array has |text_1| × |text_2| items in pair order.

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

Response Body

application/json

application/json

curl -X POST "https://example.com/v1/score" \  -H "Content-Type: application/json" \  -d '{    "model": "ms-marco",    "text_1": [      "capital of France",      "capital of Germany"    ],    "text_2": [      "Paris is the capital of France.",      "Berlin is the capital of Germany.",      "The sky is blue."    ]  }'
{  "object": "list",  "data": [    {      "object": "score",      "index": 0,      "text_1": "string",      "text_2": "string",      "score": 0.1    }  ],  "model": "string",  "usage": {    "prompt_tokens": 0,    "total_tokens": 0  }}

/rerank POST

Identical to `POST /v1/rerank` (vLLM/Jina/Cohere compatible reranker).

/pii/detect POST

Detect PII entities. Runs a token-classification model and returns entity spans with character offsets relative to the input. Provide either `text` (single) or `texts` (batch). `threshold` overrides the per-model default. `types` filters the entity types returned.