Search with a vector

Search a namespace using a query vector. It retrieves the ids of the most similar items in a namespace, along with their similarity scores.

For guidance, examples, and limits, see Search.

POST /query

Python

# pip install --upgrade pinecone
import os
from pinecone import Pinecone

pc = Pinecone(api_key=os.environ["PINECONE_API_KEY"])
# To get the host for an index, see https://docs.pinecone.io/guides/manage-data/target-an-index
index = pc.index(host="INDEX_HOST")

response = index.query(
    namespace="example-namespace",
    vector=[0.3, 0.3, 0.3, 0.3, 0.3, 0.3, 0.3, 0.3],
    filter={"genre": {"$eq": "documentary"}},
    top_k=3,
    include_values=True,
)
for match in response.matches:
    print(match.id, match.score, match.values)

200

{
  "results": [
    {
      "matches": [
        {
          "id": "example-vector-1",
          "score": 0.08,
          "values": [
            0.1,
            0.2,
            0.3,
            0.4,
            0.5,
            0.6,
            0.7,
            0.8
          ],
          "sparseValues": {
            "indices": [
              1,
              312,
              822,
              14,
              980
            ],
            "values": [
              0.1,
              0.2,
              0.3,
              0.4,
              0.5
            ]
          },
          "metadata": {
            "genre": "documentary",
            "year": 2019
          }
        }
      ],
      "namespace": "example-namespace"
    }
  ],
  "matches": [
    {
      "id": "example-vector-1",
      "score": 0.08,
      "values": [
        0.1,
        0.2,
        0.3,
        0.4,
        0.5,
        0.6,
        0.7,
        0.8
      ],
      "sparseValues": {
        "indices": [
          1,
          312,
          822,
          14,
          980
        ],
        "values": [
          0.1,
          0.2,
          0.3,
          0.4,
          0.5
        ]
      },
      "metadata": {
        "genre": "documentary",
        "year": 2019
      }
    }
  ],
  "namespace": "<string>",
  "usage": {
    "readUnits": 5,
    "egressBytes": 1024
  }
}

Python

# pip install --upgrade pinecone
import os
from pinecone import Pinecone

pc = Pinecone(api_key=os.environ["PINECONE_API_KEY"])
# To get the host for an index, see https://docs.pinecone.io/guides/manage-data/target-an-index
index = pc.index(host="INDEX_HOST")

response = index.query(
    namespace="example-namespace",
    vector=[0.3, 0.3, 0.3, 0.3, 0.3, 0.3, 0.3, 0.3],
    filter={"genre": {"$eq": "documentary"}},
    top_k=3,
    include_values=True,
)
for match in response.matches:
    print(match.id, match.score, match.values)

Authorizations

Api-Key · string · header · required

An API Key is required to call Pinecone APIs. Get yours from the console.

Headers

X-Pinecone-Api-Version · string · default: 2026-07 · required

Required date-based version header

Body

application/json

Option 1

Option 2

Option 3

The request for the query operation.

topK · integer<int64> · required

The number of results to return for each query.

Required range: 1 <= x <= 10000

Example:

10

vector · number<float>[] · required

The query vector. This should be the same length as the dimension of the index being queried. Each query request can contain only one of the parameters id or vector.

Whether vector and sparseVector may be sent together depends on the index: an index that supports only a single query predicate rejects the combination.

Required array length: 1 - 20000 elements

Example:

[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8]

namespace · string

The namespace to query.

Example:

"example-namespace"

filter · object

The filter to apply. You can use vector metadata to limit your search. See Understanding metadata.

Example:

{
  "genre": { "$in": ["comedy", "documentary", "drama"] },
  "year": { "$eq": 2019 }
}

includeValues · boolean · default: false

Indicates whether vector values are included in the response. For on-demand indexes, setting this to true may increase latency, especially with higher topK values, because vector values are retrieved from object storage. Unless you need vector values, set this to false for better performance.

Example:

true

includeMetadata · boolean · default: false

Indicates whether metadata is included in the response as well as the ids.

Example:

true

sparseVector · object

Vector sparse data. Represented as a list of indices and a list of corresponded values, which must be with the same length.

Show child attributes

id · string

The unique ID of the vector to be used as a query vector. It is mutually exclusive with both vector forms: a request naming id may not also carry vector or sparseVector.

Required string length: 1 - 512

Pattern: ^[\x01-\x7F]+$

Example:

"example-vector-1"

scanFactor · number<float>

An optimization parameter for IVF dense indexes in dedicated read node indexes. It adjusts how much of the index is scanned to find vector candidates. Range: 0.5 – 4 (default).

Keep the default (4.0) for the best search results. If query latency is too high, try lowering this value incrementally (minimum 0.5) to speed up the search at the cost of slightly lower accuracy. This parameter is only supported for dedicated (DRN) dense indexes.

Required range: 0.5 <= x <= 4

Example:

2

maxCandidates · integer<int64>

An optimization parameter that controls the maximum number of candidate dense vectors to rerank. Reranking computes exact distances to improve recall but increases query latency. Range: top_k – 100000.

Keep the default for a balance of recall and latency. Increase this value if recall is too low, or decrease it to reduce latency at the cost of accuracy. This parameter is only supported for dedicated (DRN) dense indexes.

Required range: 1 <= x <= 100000

Example:

1000

Response

200

application/json

A successful response.

The response for the query operation. These are the matches found for a particular query vector. The matches are ordered from most similar to least similar.

results · The query results for a single query vector · object[] · deprecated

DEPRECATED. The results of each query. The order is the same as QueryRequest.queries.

Show child attributes

matches · object[]

The matches for the vectors.

Show child attributes

namespace · string

The namespace that served the query: the request's namespace, or the alias's target namespace when the request named a namespace alias.

usage · object

Show child attributes