Embeddings API

The Embeddings API allows you to convert text into vector representations for semantic search and similarity matching.

Methods

embeddings.create(params)

Create embeddings for the given input text.

Parameters:

Name	Type	Required	Description
params.model	string	Yes	The embedding model to use
params.input	string | string[]	Yes	Text or array of texts to embed
params.encoding_format	'float' | 'base64'	No	Format of the embedding vectors
params.dimensions	number	No	Number of dimensions for the embeddings
params.user	string	No	Unique identifier for end-user tracking

Returns: Promise<EmbeddingResponse>

Example:

// Single text embedding
const response = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: 'The quick brown fox jumps over the lazy dog',
});

console.log(response.data[0].embedding);

// Multiple text embeddings
const response = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: [
    'First document',
    'Second document',
    'Third document',
  ],
});

response.data.forEach((item, index) => {
  console.log(`Embedding ${index}:`, item.embedding);
});

Response Format

EmbeddingResponse

Field	Type	Description
object	'list'	Object type identifier
data	EmbeddingObject[]	Array of embedding objects
model	string	The model used for embeddings
usage	EmbeddingUsage	Token usage information

EmbeddingObject

Field	Type	Description
object	'embedding'	Object type identifier
embedding	number[]	The embedding vector
index	number	Index of the embedding in the list

EmbeddingUsage

Field	Type	Description
prompt_tokens	number	Number of tokens in the input
total_tokens	number	Total tokens used

Error Handling

import { CognipeerError } from '@cognipeer/console-sdk';

try {
  const response = await client.embeddings.create({
    model: 'text-embedding-3-small',
    input: 'Hello world',
  });
  console.log(response.data);
} catch (error) {
  if (error instanceof CognipeerError) {
    console.error('Status:', error.status);
    console.error('Message:', error.message);
  }
}

Available Models

Common embedding models:

• text-embedding-3-small - Fast and efficient (1536 dimensions)
• text-embedding-3-large - Higher quality (3072 dimensions)
• text-embedding-ada-002 - Legacy model (1536 dimensions)

TIP

Use smaller models for faster processing and lower costs. Use larger models when you need higher quality embeddings for complex semantic tasks.

Use Cases

Semantic Search

// 1. Embed your documents
const docs = ['Document 1', 'Document 2', 'Document 3'];
const docEmbeddings = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: docs,
});

// 2. Embed the search query
const queryEmbedding = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: 'search query',
});

// 3. Calculate similarity and find best match
// (Use cosine similarity or store in vector database)

Text Similarity

const texts = [
  'The cat sat on the mat',
  'A dog played in the park',
];

const embeddings = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: texts,
});

// Compare embeddings to measure similarity

Best Practices

1. Batch Processing: Embed multiple texts at once to reduce API calls
2. Caching: Store embeddings to avoid re-computing for the same text
3. Normalization: Normalize embeddings before calculating similarity
4. Model Selection: Choose the right model based on your accuracy/speed requirements

Related

• Chat API - Generate text with language models
• Vectors API - Store and query embeddings in vector databases
• Vectors API - Build Retrieval Augmented Generation systems