@youcraft/recall-embeddings-openai

OpenAI embeddings provider for @youcraft/recall. Generates vector embeddings using OpenAI's embedding models.

Installation

pnpm add @youcraft/recall-embeddings-openai

Usage

import { createMemory } from '@youcraft/recall'
import { sqliteAdapter } from '@youcraft/recall-adapter-sqlite'
import { openaiEmbeddings } from '@youcraft/recall-embeddings-openai'
import { openaiExtractor } from '@youcraft/recall-extractor-openai'

const memory = createMemory({
  db: sqliteAdapter({ filename: 'memories.db' }),
  embeddings: openaiEmbeddings({ apiKey: process.env.OPENAI_API_KEY! }),
  extractor: openaiExtractor({ apiKey: process.env.OPENAI_API_KEY! }),
})

Configuration

openaiEmbeddings({
  apiKey: 'sk-...', // Required: OpenAI API key
  model: 'text-embedding-3-small', // Optional: embedding model
})

Options

| Option | Type | Default | Description | | -------- | -------- | -------------------------- | -------------------------- | | apiKey | string | Required | Your OpenAI API key | | model | string | 'text-embedding-3-small' | The embedding model to use |

Supported Models

| Model | Dimensions | Description | | ------------------------ | ---------- | ------------------------------ | | text-embedding-3-small | 1536 | Fast, cost-effective (default) | | text-embedding-3-large | 3072 | Higher quality, more expensive | | text-embedding-ada-002 | 1536 | Legacy model |

API

The provider implements the EmbeddingsProvider interface:

interface EmbeddingsProvider {
  embed(text: string): Promise<number[]>
  embedBatch(texts: string[]): Promise<number[][]>
  dimensions: number
}

embed(text)

Generate an embedding for a single text string.

const embedding = await embeddings.embed('User likes pizza')
// => [0.123, -0.456, 0.789, ...]

embedBatch(texts)

Generate embeddings for multiple texts in a single API call.

const embeddings = await embeddings.embedBatch(['User likes pizza', 'User works at Acme'])
// => [[0.123, ...], [0.456, ...]]

dimensions

The dimensionality of the embedding vectors.

console.log(embeddings.dimensions) // 1536 for text-embedding-3-small

Cost Optimization

The text-embedding-3-small model is recommended for most use cases:

• ~5x cheaper than text-embedding-3-large
• Comparable quality for similarity search tasks
• Faster response times

Use text-embedding-3-large only when you need maximum quality and can afford the extra cost.

License

MIT