Run Hugging Face open-source AI models via webhook in n8n

This workflow connects n8n to the Hugging Face Inference API, letting you run powerful open-source AI models for text generation, summarization, sentiment analysis, translation, and image generation — all fully automated, no GPU setup required. Simply POST a request and get AI-powered results back in seconds.

What's the Goal?

To give developers, agencies, and businesses a plug-and-play automation for running any Hugging Face model without managing infrastructure. Replace expensive proprietary APIs with open-source alternatives that you control.

Tasks this workflow handles out of the box:

• Text generation (GPT-style content writing)
• Summarization (condense long documents)
• Sentiment analysis (classify tone of any text)
• Translation (multilingual content)
• Image generation (text-to-image via Stable Diffusion)

Why Does It Matter?

Hugging Face hosts over 400,000 open-source models — many matching or exceeding the quality of paid APIs at a fraction of the cost. This workflow:

• Saves money: free tier available, paid plans start at $9/mo
• Gives full control: swap any model by changing one field
• Scales instantly: no GPU provisioning or DevOps needed
• Works in automation: connects to any n8n trigger or pipeline
• Produces billable output: agencies can resell AI services built on this

How It Works

Step 1 — Webhook receives the task request with input text and task type
Step 2 — Set node stores your Hugging Face API key and normalizes all inputs
Step 3 — Code node selects the right model and builds the correct API payload for the task
Step 4 — HTTP Request calls the Hugging Face Inference API with the built payload
Step 5 — Code node parses and formats the raw API response into clean structured output
Step 6 — Respond node returns the final result as JSON to the caller

Configuration Requirements

• HUGGING_FACE_API_KEY — Get free at huggingface.co/settings/tokens
• No other credentials needed for basic usage
• Optional: Google Sheets credential for logging (node already included)

Setup Guide

Step 1: Import this workflow into your n8n instance
Step 2: Open the Set API Config node and replace YOUR_HF_API_KEY with your token
Step 3: Activate the workflow
Step 4: POST to /webhook/hf-runner with your task payload
Step 5: Swap models anytime by changing the model field in your request

Sample Webhook Payload

{
"task": "summarization",
"input": "Your long text goes here...",
"model": "",
"parameters": {}
}

Supported Task Values

• text_generation
• summarization
• sentiment_analysis
• translation
• image_generation

Default Models Used

• text_generation → mistralai/Mistral-7B-Instruct-v0.2
• summarization → facebook/bart-large-cnn
• sentiment_analysis → distilbert-base-uncased-finetuned-sst-2-english
• translation → Helsinki-NLP/opus-mt-en-fr
• image_generation → stabilityai/stable-diffusion-xl-base-1.0