---
title: Embeddings
description: Generate dense vector embeddings for semantic retrieval, similarity search, and RAG pipelines.
---

<ApiEndpoint method="POST" path="/v1/embeddings" auth="api_key" />

The `/v1/embeddings` endpoint generates vector embeddings across supported embedding models (e.g. `openai/text-embedding-3-small`, `openai/text-embedding-3-large`).

<KeyInserter type="api" label="API Key Tester" placeholder="zrv_your_api_key" />

---

## Code example

<CodeGroup titles={["Python (OpenAI SDK)", "TypeScript (OpenAI SDK)", "cURL"]}>

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://api.zorveus.com/v1",
    api_key="zrv_your_api_key",
)

response = client.embeddings.create(
    model="openai/text-embedding-3-small",
    input="Vector databases index embeddings for approximate nearest neighbor search.",
)

vector = response.data[0].embedding
print(f"Generated embedding with {len(vector)} dimensions.")
```

```typescript
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.zorveus.com/v1",
  apiKey: "zrv_your_api_key",
});

async function main() {
  const response = await client.embeddings.create({
    model: "openai/text-embedding-3-small",
    input: "Vector databases index embeddings for approximate nearest neighbor search.",
  });

  console.log(`Generated embedding with ${response.data[0].embedding.length} dimensions.`);
}

main();
```

```bash
curl https://api.zorveus.com/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer zrv_your_api_key" \
  -d '{
    "model": "openai/text-embedding-3-small",
    "input": "Vector databases index embeddings for approximate nearest neighbor search."
  }'
```

</CodeGroup>

## Inputs and output

`input` can be one string or an array when the selected provider supports batching. The response contains one embedding item for each accepted input. Store the model ID and vector dimension with the vectors so that you do not compare incompatible embeddings.

Use `GET /v1/models` to find a routable embedding model. A model policy such as `openai/*` permits matching names, but it does not guarantee that a matching deployment is available.

Batch limits, token limits, dimensions, and optional encoding formats come from the selected provider and model.