Embeddings

The SambaNova Embeddings API generates vector representations (embeddings) of input text, facilitating tasks such as semantic similarity analysis, clustering, search optimization, and retrieval-augmented generation (RAG). This API enables developers to integrate advanced AI capabilities into their applications by transforming textual data into structured numerical representations.

Endpoint

The API provides an endpoint to generate embedding vectors for input text.

Request

POST https://api.sambanova.ai/v1/embeddings

For SambaStack, developers should check with their system administrator for the correct URL.

Request body parameters

Parameter	Type	Description	Required
`input`	String or array of strings	The input text to be embedded. Must not exceed the model’s token limit.	Yes
`model`	String	The model used to generate embeddings (e.g., `E5-Mistral-7B-Instruct`).	Yes

This API ensures efficient embedding generation, supporting multiple input formats while enforcing model constraints.

Example request

The following example demonstrates how to send a request to the SambaCloud Embeddings API using curl.

CURL request

curl https://api.sambanova.ai/v1/embeddings \
  -H "Authorization: Bearer <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
    "input": "The curious fox dashed through the golden field.",
    "model": "E5-Mistral-7B-Instruct"
  }'

Example response

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "embedding": [0.00333473, -0.0223934397434, ..., -0.013434322],
      "index": 0
    }
  ],
  "model": "E5-Mistral-7B-Instruct",
  "usage": {
    "prompt_tokens": 28,
    "total_tokens": 28
  }
}

Response properties

The API response consists of the following properties:

Property	Type	Description
`object`	String	The type of response, always `list`.
`data`	Array	A list of embedding objects.
`model`	String	The name of the model used to generate embeddings.
`usage`	Object	Token usage statistics for the request, including `prompt_tokens`.

Embedding object

Property	Type	Description
`object`	String	Always `embedding`.
`embedding`	Array	The embedding vector, represented as a list of floats.
`index`	Integer	The index of the embedding in the list of embeddings.

Error handling

See API error codes page for more information.

Error type	HTTP code	Description	Code
Invalid request error	400	An issue with the request parameters, e.g., model not compatible or input too long.	`invalid_request_error`
Authentication error	401	The provided API key is invalid.	`invalid_authentication`
Rate limit exceeded	429	Request quota exceeded.	`insufficient_quota`
Request timeout	408	The request timed out.	`request_timeout`

Example error response

{
  "error": {
    "message": "Model 'model_name' does not support embeddings.",
    "type": "invalid_request_error",
    "param": "model",
    "code": "model_not_compatible"
  }
}

Overview

Inference API

Using the API

Endpoint

Request

Request body parameters

Example request

CURL request

Example response

Response properties

Embedding object

Error handling

Example error response

​Endpoint

​Request

​Request body parameters

​Example request

​CURL request

​Example response

​Response properties

​Embedding object

​Error handling

​Example error response

Endpoint

Request

Request body parameters

Example request

CURL request

Example response

Response properties

Embedding object

Error handling

Example error response