Embeddings
API reference for Mixedbread's Embeddings endpoint. This documentation covers the request and response structure, supported models, and usage examples for generating embeddings using the Mixedbread API.
POST/v1/embeddings
This endpoint provides access to our embedding models. It returns embeddings for the input you provide, which can be used for various tasks such as text similarity, clustering, and more.
The endpoint is also a superset of the OpenAI embedding API. This means you can use the OpenAI API client,
pointing it to https://api.mixedbread.ai. However, note that not all Mixedbread-specific features may be
available through the OpenAI client.
- Authorization
Authorization- Type
- string
- Required or Optional
- required
- Description
This endpoint requires an API key. You can obtain one by signing up for an account on our website.
- input
input- Type
- string|string[]
- Required or Optional
- required
- Description
A string or a list of strings, where each string represents a sentence or chunk of text to be embedded.
- Between 1-256 items.
- Texts will be truncated if longer than the model's maximum sequence length
- model
model- Type
- string
- Required or Optional
- required
- Description
The model to be used for generating embeddings.
- Must be a valid model. Refer to our supported models.
- prompt
prompt- Type
- string
- Required or Optional
- optional
- Description
An optional prompt to provide context to the model. Refer to the model's documentation for more information.
- A string between 1 and 256 characters
- normalized
normalized- Type
- boolean
- Required or Optional
- optional
- Description
Option to normalize the embeddings. Defaults to true.
- dimensions
dimensions- Type
- number
- Required or Optional
- optional
- Description
The desired number of dimensions in the output vectors. Defaults to the model's maximum.
- A number between 1 and the model's maximum output dimensions
- Only applicable for Matryoshka-based models
- encoding_format
encoding_format- Type
- string|string[]
- Required or Optional
- optional
- Description
The desired format for the embeddings. Defaults to "float". If multiple formats are requested, the response will include an object with each format for each embedding.
- Options: float, float16, binary, ubinary, int8, uint8, base64
- truncation_strategy
truncation_strategy- Type
- string
- Required or Optional
- optional
- Description
The strategy for truncating input text that exceeds the model's maximum length. Defaults to "end". Setting it to "none" will result in an error if the text is too long.
- Options: start, end, none
Authentication
Request Body
Response Body
- model
model- Type
- string
- Required or Optional
- required
- Description
The embedding model used, which can be one of our hosted models or a custom fine-tuned model.
- object
object- Type
- string
- Required or Optional
- required
- Description
The type of the returned object. Always "list".
- data
data- Type
- object[]
- Required or Optional
- required
- Description
A list of the generated embeddings.
- data[x].embedding
data[x].embedding- Type
- number[]|object
- Required or Optional
- required
- Description
The vector representing the embedding, or an object with different encodings if multiple formats were requested.
- data[x].index
data[x].index- Type
- number
- Required or Optional
- required
- Description
The index of the input text corresponding to this embedding.
- data[x].object
data[x].object- Type
- number
- Required or Optional
- required
- Description
The type of the returned object. Always "embedding".
- usage
usage- Type
- object
- Required or Optional
- required
- Description
Information about API usage for this request.
- usage.prompt_tokens
usage.prompt_tokens- Type
- number
- Required or Optional
- required
- Description
The number of prompt tokens used to generate the embeddings.
- usage.total_tokens
usage.total_tokens- Type
- number
- Required or Optional
- required
- Description
The total number of tokens used to generate the embeddings.
- normalized
normalized- Type
- boolean
- Required or Optional
- required
- Description
Indicates whether the embeddings are normalized.