Embeddings

The embeddings endpoint converts text into dense vector representations you can use for semantic search, clustering, classification, and retrieval-augmented generation (RAG). It is OpenAI-compatible, so the openai SDK's embeddings.create call works without modification.

Usage#

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://opendunes.com/v1",
  apiKey: process.env.OPENDUNES_API_KEY,
});

const resp = await client.embeddings.create({
  model: "openai/text-embedding-3-large",
  input: "خوارزمي pioneered algebra in the 9th century.",
});

const vector = resp.data[0].embedding; // float32[], e.g. 3072-dimensional

Features#

• Batch inputs — pass an array of strings and receive one embedding per element in a single request.
• Multiple models — choose from several embedding models at different size/cost trade-offs; all billed in DA per million tokens.
• DA pricing — same micro-DA billing as chat: input tokens × model rate, rounded up.
• Arabic and French support — multilingual models support all three official languages of Algeria natively.

Response shape#

The response follows the OpenAI embeddings schema:

{
  "object": "list",
  "data": [
    { "object": "embedding", "index": 0, "embedding": [0.002, -0.018, ...] }
  ],
  "model": "openai/text-embedding-3-large",
  "usage": { "prompt_tokens": 12, "total_tokens": 12 }
}