Skip to main content
POST
Rerank documents
Rerank a list of documents against a query using jina-reranker-v3. Provide a non-empty query and a non-empty documents array; results come back ordered by relevance_score. Use the optional top_n to return only the most relevant documents. This pairs naturally with embeddings-based retrieval: embed and shortlist candidates, then rerank for final ordering. Served via the AIsa relay path /v1/rerank — OpenAI-compatible. Billing is token-based at $0.050 per 1M tokens; the usage.total_tokens field reports the tokens billed for each request.

Authorizations

Authorization
string
header
required

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

Body

application/json
model
string
required

Rerank model name.

Example:

"jina-reranker-v3"

query
string
required

Non-empty query string to rank documents against.

Example:

"What is the capital of France?"

documents
string[]
required

Non-empty array of document strings to rerank.

Minimum array length: 1
Example:
top_n
integer

Optional. Return only the top N most relevant documents.

Example:

2

Response

200 - application/json

Reranked results ordered by relevance.

model
string
Example:

"jina-reranker-v3"

results
object[]
usage
object