@kimio/n8n-nodes-litellm

A powerful custom node for n8n that enables seamless integration with the LiteLLM Embeddings API. Effortlessly generate AI embeddings for your workflows, connect to vector stores, and supercharge your automations with state-of-the-art language models.

🚀 Features

• Easy Integration: Plug-and-play with your LiteLLM API endpoint.
• Flexible Model Selection: Choose from available embedding models dynamically.
• Batch Processing: Efficiently embed large batches of documents.
• Customizable Options: Set batch size, embedding dimensions, and more.
• Secure Credentials: Store your API key and base URL safely using n8n's credential system.
• Ready for Vector Stores: Designed to connect directly to vector stores or AI agents in your n8n workflows.

📦 Installation

Using Docker

1. Build and copy the node to your n8n instance:

   ./install-custom-node-local.sh

2. Or, on Windows:

   install-custom-node-local.bat

Manual

1. Install dependencies and build:

   npm install
   npm run build

2. Copy the contents of the dist folder to your n8n custom nodes directory.

🛠️ Usage

1. Add the Node: In n8n, search for "LiteLLM Embeddings" and add it to your workflow.
2. Configure Credentials:
   • Set your LiteLLM API Base URL (e.g., http://localhost:4000).
   • Enter your API Key.
3. Select Model: Choose the embedding model from the dropdown.
4. Customize Options (optional):
   • Batch size
   • Embedding dimensions
   • Strip new lines
   • Override base URL
5. Connect Outputs: Link the node to a vector store, AI agent, or any downstream process.

🧩 Example Workflow

1. Input your text data.
2. Pass it to the LiteLLM Embeddings node.
3. Store or search embeddings in your favorite vector database.

🔐 Credentials

• Base URL: Your LiteLLM API endpoint.
• API Key: Your LiteLLM API key.

Credentials are securely managed via n8n's credential system.

📚 Documentation

• LiteLLM Documentation
• n8n Custom Nodes Guide

📝 License

Apache 2.0. See LICENSE.

🤝 Contributing

Pull requests and issues are welcome! Please open an issue to discuss your ideas or report bugs.

👤 Author

Felipe Kimio Nishikaku