Create Contextualized Embeddings

POST

/

v1

/

contextualizedembeddings

Create Contextualized Embeddings

curl --request POST \
  --url https://api.perplexity.ai/v1/contextualizedembeddings \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "input": [
    [
      "<string>"
    ]
  ]
}
'

{
  "object": "list",
  "data": [
    {
      "object": "list",
      "index": 123,
      "data": [
        {
          "object": "embedding",
          "index": 123,
          "embedding": "<string>"
        }
      ]
    }
  ],
  "model": "<string>",
  "usage": {
    "prompt_tokens": 123,
    "total_tokens": 123,
    "cost": {
      "input_cost": 123,
      "total_cost": 123,
      "currency": "USD"
    }
  }
}

{
  "detail": [
    {
      "loc": [
        "<string>"
      ],
      "msg": "<string>",
      "type": "<string>"
    }
  ]
}

Authorization

string

header

required

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

application/json

Request body for creating contextualized embeddings

input

string[][]

required

Nested array structure where each inner array contains chunks from a single document. Chunks within the same document are encoded with document-level context awareness. Maximum 512 documents. Total chunks across all documents must not exceed 16,000. Total tokens per document must not exceed 32K. All chunks in a single request must not exceed 120,000 tokens combined. Empty strings are not allowed.

Required array length: 1 - 512 elements

Minimum array length: 1

Minimum string length: 1

model

enum<string>

required

The contextualized embedding model to use

Available options:

pplx-embed-context-v1-0.6b,

pplx-embed-context-v1-4b

dimensions

integer

Number of dimensions for output embeddings (Matryoshka). Range: 128-1024 for pplx-embed-context-v1-0.6b, 128-2560 for pplx-embed-context-v1-4b. Defaults to full dimensions (1024 or 2560).

Required range: 128 <= x <= 2560

encoding_format

enum<string>

default:base64_int8

Output encoding format for embeddings. base64_int8 returns base64-encoded signed int8 values. base64_binary returns base64-encoded packed binary (1 bit per dimension).

Available options:

base64_int8,

base64_binary

Successful Response

Response body for contextualized embeddings request

object

string

The object type

Example:

"list"

data

Contextualized Embedding Object · object[]

List of contextualized embedding objects

model

string

The model used to generate embeddings

usage

Embeddings Usage · object

Token usage for the embeddings request

Was this page helpful?

⌘I