Rerank

POST /v1/rerank

Scores and reorders a set of documents by relevance to a query (Jina / Cohere Rerank compatible). Commonly used to refine RAG recall.

RequestREQUEST

POSThttps://test.kailin.com.cn/v1/rerank

API KeyStored only in your browser, never uploaded

Body (JSON){ "model": "jina-reranker-v2-base-multilingual", "query": "What is vector search", "documents": [ "Vector search computes semantic similarity over embedding vectors", "The weather is nice today", "RAG systems typically retrieve first, then rerank" ], "top_n": 2 }

ResponseRESPONSE

Fill in the request and hit Send — the response shows up here

Request

curl https://test.kailin.com.cn/v1/rerank \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-your-token" \
  -d '{
    "model": "jina-reranker-v2-base-multilingual",
    "query": "What is vector search",
    "documents": [
      "Vector search computes semantic similarity over embedding vectors",
      "The weather is nice today",
      "RAG systems typically retrieve first, then rerank"
    ],
    "top_n": 2
  }'

Request Body

Parameter	Type	Required	Description
model	string	Yes	Rerank model — see the available list
query	string	Yes	Query text
documents	array	Yes	Documents as strings or { "text": "..." } objects
top_n	integer	No	Return only the top N results, default all
return_documents	boolean	No	Include the original text in results, default false

Response

{
  "results": [
    {
      "index": 0,
      "relevance_score": 0.93,
      "document": { "text": "Vector search computes semantic similarity over embedding vectors" }
    },
    {
      "index": 2,
      "relevance_score": 0.71,
      "document": { "text": "RAG systems typically retrieve first, then rerank" }
    }
  ],
  "usage": {
    "total_tokens": 120
  }
}

Field	Description
results[].index	Original index in the input documents array
results[].relevance_score	Relevance score, higher is more relevant
usage.total_tokens	Billed tokens (may be absent on some channels)

tip

Typical pattern: vector recall top 50 → rerank to top 5 → feed into the chat model as context.