Rerank Documents
jina.rerank
ActionAI & modelsconnection required
Calls the Jina AI API with the connection's credentials.
Score a list of documents against a query with a Jina reranker and return them ordered by relevance.
Connection
- Required
- Yes — the node will not save without one
- Auth type
- Bearer token
- Providers
- Jina AI
- Secret shape
- the API key, or {"api_key": "…"}
- Config it needs
- base_url, default_embedding_model, default_reranker_model
- Resolved by
- config.connection_id, else node.connection_id
Configuration
7 fields, 2 required. A dotted path is a nested object and name[] an array of objects — the editor renders the same tree.
| Field | Type | Required | Default | Constraints | Example |
|---|---|---|---|---|---|
query The query the documents are ranked against. | string | Yes | — | — | how do I rotate an API key |
documents Documents to rank: strings, or objects such as {"text": "…"} and {"image": "https://…"} for the multimodal reranker. | array | Yes | — | — | ["rotate your key in the dashboard","unrelated text"] |
model Reranker model, for example jina-reranker-v3.5 or jina-reranker-v2-base-multilingual. Falls back to the connection's default. | string | No | — | — | jina-reranker-v3.5 |
top_n Return only this many of the best-scoring documents. | number | No | — | ≥ 1 | — |
return_documents Include each document's text in the results, not just its index and score. | boolean | No | — | — | — |
max_doc_length Tokens to keep per document, on the jina-reranker-v3 family. | number | No | — | 1–8192 | — |
return_embeddings Also return each document's embedding, on the jina-reranker-v3 family. | boolean | No | — | — | — |
Example configuration
Every value below is the contract’s own example or default — nothing invented.
{
"documents": [
"rotate your key in the dashboard",
"unrelated text"
],
"model": "jina-reranker-v3.5",
"query": "how do I rotate an API key"
}Output
One shape.
result
| Field | Type | Example |
|---|---|---|
model The reranker that scored the documents. | string | — |
results Documents ordered by relevance, best first. | array | — |
results[].index The document's position in the input. | number | — |
results[].relevance_score Relevance score; higher is better. | number | — |
results[].document The document, when return_documents was on. | any | — |
count How many results came back. | number | — |
top_index Input index of the best-scoring document. | number | — |
top_score Relevance score of that best document. | number | — |
usage Token usage Jina billed for this call. | object | — |
{
"count": 0,
"model": "",
"results": [
{
"document": null,
"index": 0,
"relevance_score": 0
}
],
"top_index": 0,
"top_score": 0,
"usage": {}
}See Rerank Documents run in a workflow.
In a technical demo we add the Jina AI connection, drop the node on the canvas and let Workflow Health pre-flight it before the first real run.