n8n Workflow Basics

Purpose

Document the core workflow concepts that became important once I started connecting n8n to Ollama and treating automation as a practical system instead of a visual diagram.

Key Idea

n8n is most useful when I treat it as a data orchestration tool. The main job is not "AI" or "automation" in the abstract. The main job is moving inputs through a sequence of clear steps and returning something useful on the other side.

Explanation

What n8n Is

n8n is an automation platform that allows different tools and services to communicate through workflows.

It can connect:

• APIs
• Databases
• AI models
• Web applications
• Email systems
• Spreadsheets
• Custom scripts

without requiring a full custom application for every integration.

The name n8n stands for Node Automation.

Why I Started Using It

I wanted a way to connect local AI models with other tools without writing an entire application first.

The main goals were:

• learn automation concepts
• build AI workflows
• experiment with local AI
• reduce repetitive work
• create reusable systems

n8n provided a visual way to understand how data moves between services.

Mental Model

The simplest mental model is an assembly line.

• workflow = assembly line
• node = machine
• data = package
• trigger = start button
• output = finished product

Each machine performs a specific task and passes the result to the next machine.

Workflow

A workflow is a sequence of connected steps.

Workflows define:

• what happens
• when it happens
• what data moves between steps

Node

A node performs a specific action.

Examples:

• HTTP Request
• Edit Fields
• Schedule Trigger
• Code
• AI Agent

Every workflow is made up of nodes.

Trigger

A trigger starts a workflow.

Examples:

• Manual Trigger
• Schedule
• Webhook
• Form submission
• Email received

Without a trigger, a workflow does not start.

Data Flow

Data moves from node to node.

Each node:

1. Receives data
2. Processes data
3. Sends data forward

Understanding data flow is one of the most important skills in n8n, because most debugging comes down to figuring out what was passed in, how it changed, and what the next node expected.

Practical Use

My first useful workflow was simple: send text to a locally running Ollama model and receive a response.

Workflow:

Manual Trigger
    ↓
Edit Fields
    ↓
HTTP Request
    ↓
Response

This successfully connected n8n to a local AI model running through Ollama.

Connecting n8n To Ollama

Ollama exposes an API locally.

Example endpoint:

http://ollama:11434/api/generate

In Docker Compose, containers can communicate using service names instead of localhost.

This allowed the n8n container to communicate directly with the Ollama container.

Common beginner mistakes showed up quickly:

• Invalid JSON: One missing quote or comma can break an API request. Always validate JSON before testing.
• Incorrect data types: A string like "false" is not the same as a boolean false.
• Testing too much at once: When troubleshooting, test one node at a time and verify input, output, and data format before moving forward.
• Skipping logs: Logs often explain exactly what failed if I read the error message carefully before changing multiple things at once.

Incorrect:

{
  "stream": "false"
}

Correct:

{
  "stream": false
}

Boolean values should not be wrapped in quotation marks.

Related Experiments

• Local AI Lab 001: First n8n to Ollama Test
• Local AI Lab 002: First n8n to Ollama Workflow
• Local AI Lab 003: Summarizing Qurio Notes with Ollama and n8n

Takeaways

• n8n became easier once I stopped thinking about it as "magic automation" and started thinking about inputs, transformations, and outputs.
• The most important beginner skill is tracing data flow between nodes.
• Automation is mostly moving data between systems cleanly enough that the next step can use the result.
• Small workflows are easier to debug than broad ones.
• Local AI fits naturally into n8n because Ollama behaves like any other HTTP service.