Skip to content
Docs

OpenAI-Compatible Embeddings API with AI Gateway

Generate vector embeddings from input text using the OpenAI-compatible /embeddings endpoint, for semantic search, similarity matching, and retrieval-augmented generation (RAG).

For an overview of embedding models and the AI SDK (embed, embedMany), see Embeddings. This page covers the OpenAI-compatible REST endpoint.

Endpoint
POST /embeddings
Example request
embeddings.ts
import OpenAI from 'openai';
 
const apiKey = process.env.AI_GATEWAY_API_KEY || process.env.VERCEL_OIDC_TOKEN;
 
const openai = new OpenAI({
  apiKey,
  baseURL: 'https://ai-gateway.vercel.sh/v1',
});
 
const response = await openai.embeddings.create({
  model: 'openai/text-embedding-3-small',
  input: 'Sunny day at the beach',
});
 
console.log(response.data[0].embedding);
embeddings.py
import os
from openai import OpenAI
 
api_key = os.getenv("AI_GATEWAY_API_KEY") or os.getenv("VERCEL_OIDC_TOKEN")
 
client = OpenAI(
    api_key=api_key,
    base_url="https://ai-gateway.vercel.sh/v1",
)
 
response = client.embeddings.create(
    model="openai/text-embedding-3-small",
    input="Sunny day at the beach",
)
 
print(response.data[0].embedding)
embeddings.sh
curl -X POST "https://ai-gateway.vercel.sh/v1/embeddings" \
  -H "Authorization: Bearer $AI_GATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/text-embedding-3-small",
    "input": "Sunny day at the beach"
  }'
Response format

The following response truncates the embedding vector for readability:

{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [-0.0038, 0.021]
    }
  ],
  "model": "openai/text-embedding-3-small",
  "usage": {
    "prompt_tokens": 6,
    "total_tokens": 6
  },
  "providerMetadata": {
    "gateway": {
      "routing": {},
      "cost": "0.00000012"
    }
  }
}
Dimensions parameter

You can set the root-level dimensions field (from the OpenAI Embeddings API spec) and the gateway will auto-map it to each provider's expected field; providerOptions.[provider] still passes through as-is and isn't required for dimensions to work.

embeddings-dimensions.ts
import OpenAI from 'openai';
 
const openai = new OpenAI({
  apiKey: process.env.AI_GATEWAY_API_KEY,
  baseURL: 'https://ai-gateway.vercel.sh/v1',
});
 
const response = await openai.embeddings.create({
  model: 'openai/text-embedding-3-small',
  input: 'Sunny day at the beach',
  dimensions: 768,
});
embeddings-dimensions.py
import os
from openai import OpenAI
 
client = OpenAI(
    api_key=os.getenv('AI_GATEWAY_API_KEY'),
    base_url='https://ai-gateway.vercel.sh/v1'
)
 
response = client.embeddings.create(
    model='openai/text-embedding-3-small',
    input='Sunny day at the beach',
    dimensions=768,
)
embeddings-dimensions.sh
curl -X POST "https://ai-gateway.vercel.sh/v1/embeddings" \
  -H "Authorization: Bearer $AI_GATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/text-embedding-3-small",
    "input": "Sunny day at the beach",
    "dimensions": 768
  }'
Last updated September 8, 2026
Previous
Advanced
Next
Image Generation

Was this helpful?

supported.