Use an open-source LLM (via HuggingFace)

Nodes

6607a47e-4139-412e-a10a-5c4907526331

Created by

n8n8n Team

Last edited 38 days ago

This workflow demonstrates how to connect an open-source model to a Basic LLM node.

The workflow is triggered when a new manual chat message appears. The message is then run through a Language Model Chain that is set up to process text with a specific prompt to guide the model's responses.

Note that open-source LLMs with a small number of parameters require slightly different prompting with more guidance to the model.

You can change the default Mistral-7B-Instruct-v0.1 model to any other LLM supported by HuggingFace. You can also connect other nodes, such as Ollama.

Note that to use this template, you need to be on n8n version 1.19.4 or later.

New to n8n?

Need help building new n8n workflows? Process automation for you or your company will save you time and money, and it's completely free!