> For the complete documentation index, see [llms.txt](https://dnotifier.gitbook.io/product-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://dnotifier.gitbook.io/product-docs/integrations/embedding-models.md).

# Embedding models

Set `embedding` on the client, or pass it on a knowledge-base call. Leave it unset to use the embedding model on your subscription.

```js
notifier.serverConfig = {
  embedding: {
    provider: "open_ai",
    model: "text-embedding-3-small",
  },
};
```

`embeddingConfig` is another name for the same object. Include `dimensions` when the model should use a specific size.

| Provider     | `provider`     | Model                                                          |
| ------------ | -------------- | -------------------------------------------------------------- |
| OpenAI       | `open_ai`      | `text-embedding-3-small`                                       |
| Azure AI     | `azure_ai`     | Your deployment. Set `deploymentName` when `model` is omitted. |
| Gemini       | `gemini`       | `text-embedding-004`                                           |
| AWS Bedrock  | `bedrock`      | `amazon.titan-embed-text-v2:0`                                 |
| Hugging Face | `hugging_face` | `sentence-transformers/all-MiniLM-L6-v2`                       |
| Ollama       | `ollama`       | `nomic-embed-text`                                             |
| Cohere       | `cohere`       | `embed-english-v3.0`                                           |
| Voyage       | `voyage`       | `voyage-3`                                                     |
| Mistral      | `mistral`      | `mistral-embed`                                                |
| Jina         | `jina`         | `jina-embeddings-v3`                                           |

These names are also accepted:

| Instead of     | You can pass                           |
| -------------- | -------------------------------------- |
| `azure_ai`     | `azure`, `azure_openai`                |
| `gemini`       | `google`                               |
| `bedrock`      | `aws`, `aws_bedrock`, `amazon_bedrock` |
| `hugging_face` | `huggingface`                          |
| `voyage`       | `voyageai`, `voyage_ai`                |
| `mistral`      | `mistral_ai`                           |
| `jina`         | `jina_ai`                              |

Any other `provider` is an OpenAI-compatible endpoint. Provide `apiKey` and `baseUrl`.

Ollama does not use an API key and uses `nomic-embed-text` instead of the subscription model. Bedrock `region` defaults to `us-east-1`.

An `embedding` on one call overrides the client setting for that call. The vector store stays as you set it.

→ [How to set an embedding model](https://dnotifier.gitbook.io/product-docs/ai/vector-stores.md) · [Vector stores](https://dnotifier.gitbook.io/product-docs/integrations/vector-stores.md)
