Create Embeddings

Generate embeddings for a list of texts. Use these embeddings for semantic search, clustering, and other machine learning applications.

POST /v1/embeddings

Create Embeddings

cURL

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

200

{
  "object": "list",
  "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"
    }
  }
}

Adding embeddings to an app you already have? The quickstart’s existing-project recipe puts the client in one module, indexes the records your app already stores, and ranks from storage. To decode embeddings into float32 vectors before you write them to a vector database, see Prepare vectors for a vector database.

Authorizations

Authorization · string · header · required

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

Body

application/json

Request body for creating embeddings

input · string string[] · required

Input text to embed, encoded as a string or array of strings. Maximum 512 texts per request. Each input must not exceed 32K tokens. All inputs in a single request must not exceed 120,000 tokens combined. Empty strings are not allowed.

Minimum string length: 1

model · enum<string> · required

The embedding model to use

Available options: pplx-embed-v1-0.6b, pplx-embed-v1-4b

dimensions · integer

Number of dimensions for output embeddings (Matryoshka). Range: 128-1024 for pplx-embed-v1-0.6b, 128-2560 for pplx-embed-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

Response

200

application/json

Successful Response

Response body for embeddings request

object · string

The object type

Example:

"list"

data · Embedding Object · object[]

List of embedding objects

Show child attributes

model · string

The model used to generate embeddings

usage · Embeddings Usage · object

Token usage for the embeddings request

Show child attributes