How can I integrate a non-English embeddings model (multilingual-e5) into n8n workflows?

Below is a summary of your options when you want to incorporate a non-English embedding model (like multilingual-e5) into your n8n workflows:

Use the HTTP Request Node

Since there’s no native support for Pinecone’s embedding service (which supports multilingual-e5) in n8n’s built-in embeddings nodes, you can use the generic HTTP Request node to call the API directly. In this approach, you would:

• Build your API request manually: Configure the HTTP Request node to send the necessary payload (your text) to the Pinecone embeddings endpoint.
• Process the response: Use the output from the HTTP Request node as input for your vector database nodes.

Create a Custom Node or Integration

If you plan to use this capability frequently, consider developing a custom node or integration for Pinecone that wraps their API. This way, you can expose the multilingual-e5 functionality as a dedicated node inside n8n, streamlining integration with your vector database nodes.

Use a Function Node for Custom Code

For a less formal solution, you can write custom code within a Function node. This approach allows you to handle API calls to Pinecone or any similar provider and process the response directly in JavaScript. The output from your Function node can then be used by subsequent nodes in your workflow.

Final Considerations

• Authentication and credentials: Make sure to handle API credentials securely, using either the HTTP Request node’s credentials options or custom credential management if you build a custom node.
• Error Handling: Whichever option you choose, implement appropriate error handling—especially when calling external APIs—to ensure your workflow is robust.

By using one of these approaches, you can successfully integrate a non-English embedding model like multilingual-e5 into your n8n workflows even though native support isn’t available at the moment.