Tracing n8n

n8n is a fair-code workflow automation platform that lets you connect services and build AI-powered workflows visually. MLflow integrates with n8n via the n8n-nodes-mlflow community node — an OpenAI-compatible chat model node with built-in MLflow tracing. Every LLM request and response is automatically logged as a trace, with multi-turn conversations grouped by session in the MLflow UI.

Features

• OpenAI-compatible — works with OpenAI, LiteLLM, LocalAI, Azure OpenAI, and any OpenAI-compatible endpoint
• Automatic tracing — every LLM call is logged as an MLflow trace with inputs, outputs, and latency
• Session grouping — traces are linked by sessionId for multi-turn conversation tracking
• Custom metadata — attach sessionId, userId, and arbitrary JSON to every trace span
• Databricks support — compatible with both self-hosted and Databricks-managed MLflow

Setup

1. Installation

Option 1: Community Nodes UI (Recommended, n8n v0.187+)

1. Go to Settings → Community Nodes
2. Click Install
3. Enter n8n-nodes-mlflow as the npm package name
4. Agree to the risks of using community nodes and click Install

Option 2: Docker

A preconfigured Docker setup is available in the docker/ directory:

bash

# Build the image
docker build --progress=plain -f docker/Dockerfile -t n8n-nodes-mlflow .

# Run the container
docker run -it --rm \
  --name n8n-mlflow \
  -p 5678:5678 \
  -e GENERIC_TIMEZONE="America/New_York" \
  -e TZ="America/New_York" \
  -e N8N_ENFORCE_SETTINGS_FILE_PERMISSIONS=true \
  -e N8N_RUNNERS_ENABLED=true \
  -v n8n_data:/home/node/.n8n \
  n8n-nodes-mlflow:latest

Access n8n at http://localhost:5678.

Option 3: Manual Installation

bash

cd ~/.n8n
npm install n8n-nodes-mlflow
n8n start

2. Credentials

Configure credentials in n8n for both the LLM endpoint and MLflow.

OpenAI Settings

Field	Required	Description	Example
OpenAI API Key	Yes	API key for your OpenAI-compatible endpoint	sk-abc123...
OpenAI Base URL	No	Override the default endpoint base URL	https://api.openai.com/v1

MLflow Settings

Field	Required	Description	Example
MLflow Tracking URI	Yes	Your MLflow tracking server URI	http://localhost:5000 or databricks
MLflow Experiment ID	Yes	Experiment ID where traces will be logged	1234567890

:::tip Databricks users Set the Tracking URI to databricks and provide your Databricks personal access token as the MLflow Tracking Token. :::

3. Node Configuration

Add the OpenAI MLflow node to your workflow and configure the MLflow tracing fields:

Field	Type	Description
sessionId	string	Groups related traces into one chat session (mlflow.trace.session)
userId	string	Identifies the end user making the request (mlflow.trace.user)
customMetadata	object	Arbitrary JSON attached to the trace as span metadata

Workflow Setup

A typical n8n workflow using this node:

Chat Example

Sending a chat message through n8n:

Viewing Traces in MLflow

After running a workflow, open the MLflow UI and navigate to the Traces tab. Each LLM call appears as a trace with full inputs, outputs, and latency:

Enable "Group by session" to see all turns of a conversation grouped under one session:

Compatibility

• n8n version 1.0.0 or later
• Node.js >= 20.15
• OpenAI official API, LiteLLM, LocalAI, Azure OpenAI, or any OpenAI-compatible endpoint
• Self-hosted MLflow and Databricks-managed MLflow

Resources

• n8n-nodes-mlflow on npm
• mlflow/mlflow-n8n-nodes on GitHub
• n8n Community Nodes documentation
• MLflow Tracing documentation

Next steps

Track User Feedback

Record user feedback on traces for tracking user satisfaction.

Learn about feedback →

Manage Prompts

Learn how to manage prompts with MLflow's prompt registry.

Manage prompts →

Evaluate Traces

Evaluate traces with LLM judges to understand and improve your AI application's behavior.

Evaluate traces →