n8n-nodes-openrouter-embeddings

An n8n community sub-node that lets you generate text embeddings through OpenRouter's unified API. Access embedding models from OpenAI, Cohere, Google, and many other providers with a single credential.

Features

• Multi-provider access – use OpenAI, Cohere, Google, and other embedding models through one API
• Dynamic model list – models are loaded directly from the OpenRouter API so you always see the latest options
• Batch processing – configurable batch size for efficient bulk embedding
• Drop-in sub-node – works with every n8n vector store and AI chain node
• Simple auth – only requires an OpenRouter API key
• Debug logging – set N8N_NODES_DEBUG=1 for detailed request/response logs

Installation

Community Node (Recommended)

1. In n8n, go to Settings → Community Nodes
2. Search for n8n-nodes-openrouter-embeddings
3. Click Install

Manual Installation

npm install n8n-nodes-openrouter-embeddings

Setup

Prerequisites

1. An OpenRouter account
2. An API key – create one at https://openrouter.ai/keys

Authentication

1. In n8n, go to Credentials → New Credential
2. Search for OpenRouter API
3. Paste your API key and save

Usage

This is a sub-node that supplies embeddings to other n8n AI nodes (vector stores, chains, agents).

Using with Vector Stores

1. Add a vector store node to your workflow (e.g., Pinecone, Qdrant, Supabase)
2. Connect Embeddings OpenRouter to the embeddings input
3. Select your OpenRouter API credential
4. Pick a model from the dropdown (e.g., openai/text-embedding-3-small)
5. Configure options as needed

Example Workflow

[Document Loader] → [Vector Store] ← [Embeddings OpenRouter]
                          ↓
                    [AI Agent / Chain]

Configuration

| Parameter | Description | |-----------|-------------| | Model | Embedding model to use. The dropdown is populated dynamically from the OpenRouter API. Default: openai/text-embedding-3-small. | | Strip New Lines | Remove \n characters before embedding (default: true). | | Batch Size | Maximum texts per API call (default: 100). Lower this if you hit rate limits. |

Popular Models

Browse all available models at https://openrouter.ai/models?output_modalities=embeddings.

Some commonly used models:

| Model ID | Provider | Notes | |----------|----------|-------| | openai/text-embedding-3-small | OpenAI | Fast, cost-effective, 1 536 dimensions | | openai/text-embedding-3-large | OpenAI | Higher quality, 3 072 dimensions | | cohere/embed-multilingual-v3.0 | Cohere | Excellent multilingual support | | google/text-embedding-004 | Google | 768 dimensions |

Use Cases

• RAG (Retrieval-Augmented Generation) – embed documents and queries for context retrieval
• Semantic Search – find relevant documents by meaning instead of keywords
• Document Similarity – compare and cluster documents in a vector store
• Classification & Clustering – group or categorize text by embedding proximity

Compatible Nodes

Works with any n8n node that accepts an embeddings sub-node, including:

• Simple Vector Store
• Pinecone Vector Store
• Qdrant Vector Store
• Supabase Vector Store
• PGVector Vector Store
• Milvus Vector Store
• MongoDB Atlas Vector Store
• Zep Vector Store
• Question and Answer Chain
• AI Agent nodes

Troubleshooting

| Error | Cause | Fix | |-------|-------|-----| | 401 Unauthorized | Invalid or missing API key | Verify your key at https://openrouter.ai/keys | | 402 Payment Required | Insufficient credits | Top up your OpenRouter balance | | 404 Not Found | Model doesn't exist or isn't an embedding model | Check the model ID and the models page | | 429 Too Many Requests | Rate limit exceeded | Lower the batch size or add delays | | 529 Provider Overloaded | Upstream provider is overloaded | Retry later or try a different model |

API Reference

This node calls the OpenRouter Embeddings API:

POST https://openrouter.ai/api/v1/embeddings

Full API docs: https://openrouter.ai/docs/api/reference/embeddings

Contributing

Contributions are welcome! Please open a Pull Request or Issue.

License

MIT

Changelog

0.1.0

• Initial release
• OpenRouter embeddings support with dynamic model loading
• Batch processing with configurable batch size
• Strip new-lines option
• Custom credential with connection test