🚀 Build Local AI Automation Workflows with Ollama + n8n: A Complete Guide

D: The combination of Ollama (for running local LLMs) and n8n (a powerful workflow automation tool) opens up exciting possibilities for private, self-hosted AI automation. Let’s dive into how to connect these tools and build intelligent workflows—100% offline if needed!

🔧 Prerequisites

Before we start, make sure you have:

• Ollama installed (Download here) – Run models like llama3, mistral, or phi3 locally.
• n8n set up (Installation guide) – Self-hosted or cloud version.
• Basic familiarity with APIs and JSON.

🌟 Step 1: Run Ollama & Enable API

Ollama’s REST API lets n8n interact with your local AI. Start Ollama with:

ollama serve

By default, the API runs at http://localhost:11434.

🔹 Test the API

Try a curl command to confirm it works:

curl http://localhost:11434/api/generate -d '{
  "model": "llama3",
  "prompt": "Hello, how are you?"
}'

You should see a streamed AI response.

🤖 Step 2: Connect Ollama to n8n

In n8n, use the HTTP Request node to call Ollama’s API.

📌 Example Workflow: AI-Powered Content Summarizer

1. Trigger: Use an n8n Schedule Trigger (e.g., hourly).
2. HTTP Request Node:
   • Method: POST
   • URL: http://localhost:11434/api/generate
   • Headers: Content-Type: application/json
   • Body (JSON):

     {
     "model": "mistral",
     "prompt": "Summarize this text: {{$node["Webhook"].json["input_text"]}}"
     }

3. Output: Save the AI response to a file/database or send it via email/Slack.

🛠 Advanced Use Cases

1️⃣ Automated Document Processing

• Flow:
  • Watch a folder (via n8n File Trigger).
  • Extract text (OCR if needed).
  • Send to Ollama for summarization/translation.
  • Output results to Notion or Google Docs.

2️⃣ AI Chatbot with Slack

• Use n8n’s Slack Trigger to detect messages.
• Route queries to Ollama (llama3 for creative answers, phi3 for facts).
• Post replies automatically.

3️⃣ Data Enrichment Pipeline

• Fetch raw data (e.g., CRM contacts).
• Use Ollama to categorize sentiment or generate tags.
• Update your database.

⚠️ Tips & Troubleshooting

• Performance: Larger models (e.g., llama3-70b) need significant RAM. Start with mistral or tinyllama for lightweight tasks.
• Streaming: Ollama’s API streams responses. Use "stream": false if you prefer a complete reply.
• Security: If exposing n8n/Ollama online, add authentication (e.g., n8n user management, Ollama behind a reverse proxy).

🔥 Why This Combo Rocks

• Privacy: No data leaves your machine.
• Flexibility: Chain AI with 300+ n8n integrations (Discord, Airtable, etc.).
• Cost-Efficient: Avoid GPT-4 API fees for simple tasks.

📚 Resources

• Ollama Model Library
• n8n Nodes Documentation

💡 Pro Tip: Combine with Tavily (local web search) for AI agents that research autonomously!

Ready to build? Start small—try summarizing emails or generating SQL queries, then scale up! 🚀