Frosty AI n8n Integration Guide 🚀

Complete guide for using Frosty AI inside your self-hosted n8n workflows.

💡 What is Frosty AI?

Frosty AI is an AI model routing and observability platform that helps you optimize AI model usage across multiple providers. With Frosty, you can:

  • Route prompts across OpenAI, Anthropic, Mistral, and more

  • Automatically select models based on cost, performance, or fallback

  • Track usage, costs, latency, and success rates

  • Reduce vendor lock-in and improve reliability with failover logic

This guide walks you through how to use the Frosty AI community node in n8n.

🔗 Getting Started with Frosty AI in n8n

To use the Frosty AI n8n integration, you need:

✅ A Frosty AI account → Sign up at console.gofrosty.ai
✅ A router set up in the Frosty console → Use the Quick Start Wizard to create your first router
✅ A self-hosted (locally running) instance of n8n → Follow the n8n self-hosting docs to get started

🛠️ Installing Frosty AI in n8n

⚠️ The Frosty AI node works with self-hosted n8n instances only. Community nodes are not currently supported in n8n Cloud.

Step 2: Install the Community Node

In your self-hosted n8n environment, run:

n8n install n8n-nodes-frosty-ai

Then restart n8n and open the editor.

🤖 Using the Frosty AI Node in a Workflow

Example: Route a prompt and return an AI response

This basic workflow sends a user prompt through your Frosty router and returns the response.

Workflow steps:

  1. Webhook Trigger – Accepts a POST request with a prompt

  2. Frosty AI Node – Sends the prompt to your router

  3. Respond to Webhook – Returns the AI-generated reply

[Webhook Trigger] → [Frosty AI Node] → [Respond to Webhook]

🔐 Setting Up Credentials

When prompted by the Frosty AI node:

  1. Click Add New to create a credential

  2. Paste in your Router ID and Router Key

  3. Click Save

n8n will securely store your credentials for future use.

📝 Inputs (Frosty AI Chat Operation)

Parameter

Description

Required

prompt

The AI prompt to send

✅ Yes

rule

Routing logic: cost, performance, or none

✅ Yes

📤 Outputs (Frosty AI Node)

Field

Description

trace_id

Unique request identifier

model

AI model used

provider

AI provider (e.g., OpenAI, Anthropic)

response

AI-generated response

cost

Estimated cost for the request

total_tokens

Total tokens used (input + output)

success

True if request was successful

🛑 Error Handling

Frosty AI returns structured errors to help with workflow reliability:

Status Code

Meaning

401

Unauthorized – Check Router ID/Key

403

Access Denied – Trial expired or access removed

429

Rate Limit – Too many requests

502

No Provider Available – All models failed

500

Internal Error – Unknown server issue

Use n8n’s built-in error handling tools to retry, fallback, or log failed executions.

💡 Example Use Cases

  • 🤖 Smart auto-replies in forms and webhooks

  • 💰 Cost-based routing for model selection

  • 🛡️ Failover across multiple LLM providers

  • 📊 Log and analyze prompt behavior over time

  • 🧠 A/B test different AI providers

🙌 Need Help?

For questions or support, reach out to us:

📩 Email: support@gofrosty.ai
🔗 Console: https://console.gofrosty.ai
📚 Docs: https://docs.gofrosty.ai

✨ Start building smarter AI workflows with Frosty AI + n8n today!

Updated on
Make Integration Guide 🚀