Phase 2: n8n for AI Automation · 50 min · n8n · JavaScript

Sub-Workflows and Code Nodes — Advanced n8n Patterns

Modular workflows are maintainable workflows.

Hiring signal: Sub-workflows and Code nodes are what separate n8n beginners from n8n power users. In technical interviews, being able to explain when to use a sub-workflow vs a Code node, and write a Code node that transforms JSON data, demonstrates real n8n production experience.

What you will learn

Build n8n sub-workflows for reusable logic and modular automation design
Use Code nodes to write custom JavaScript for data transformation, API calls, and conditional routing
Decide when to use a Code node vs a built-in node
Apply common Code node patterns: JSON parsing, date formatting, array filtering, HTTP requests

The Problem

A marketing team builds an n8n workflow that handles lead processing: receive webhook → enrich with Clearbit → score with AI → route by score → add to CRM → send welcome email. It works. Then they add a second workflow for customer support tickets: receive email → classify with AI → route by category → create ticket → notify agent. Both workflows need the same AI classification logic and the same CRM update pattern. Instead of reusing, they copy-paste the nodes. Three months later, the AI prompt needs updating — they change it in one workflow, forget the other, and leads get misclassified.

This is the modularity problem. Without sub-workflows and Code nodes, n8n workflows become monolithic, duplicated, and unmaintainable. Sub-workflows let you build reusable logic blocks. Code nodes let you write custom JavaScript when no built-in node fits.

Treat workflows like code, not scripts
Every principle from software engineering applies: DRY (don't repeat yourself), single responsibility, separation of concerns. Sub-workflows are your functions. Code nodes are your escape hatch when built-in nodes can't express the logic.

The Concept

Sub-Workflows: Functions for n8n

A sub-workflow is a workflow that gets called by another workflow. The parent passes data in, the sub-workflow processes it, and returns results.

┌─────────────────────────────────────────────────────────┐
│  PARENT WORKFLOW                                        │
│                                                         │
│  [Trigger] → [Transform] → [Execute Sub-Workflow] → ... │
│                              │                          │
│                              ▼                          │
│                    ┌──────────────────────┐             │
│                    │  SUB-WORKFLOW        │             │
│                    │                      │             │
│                    │  [Input] → [Process] │             │
│                    │  → [Return Value]    │             │
│                    └──────────────────────┘             │
│                              │                          │
│                              ▼                          │
│                         [Continue]                      │
└─────────────────────────────────────────────────────────┘

Property	Main Workflow	Sub-Workflow
Trigger	Webhook, schedule, manual	Execute Workflow trigger (called by parent)
Input	External event	Data from parent workflow
Output	Final result	Return value to parent
Reusable	Usually no	Yes — called from multiple parents

Code Nodes: The Escape Hatch

Code nodes let you write JavaScript (or Python) directly inside n8n. Use them when:

Data transformation that's too complex for the Set node
Conditional routing with logic beyond IF/Switch nodes
Custom API calls that need header manipulation or signing
Array manipulation: filtering, mapping, grouping, deduplication
Date/time arithmetic beyond what the Date node supports

You have three workflows that all need to validate and normalize phone numbers. What's the best approach?

A sub-workflow centralizes the logic. When the validation rules change (e.g., adding a new country code format), you update one sub-workflow and all three parent workflows automatically use the new logic. Duplicating Code nodes means three places to update — and you'll forget one.

When to Use a Code Node vs a Built-in Node

Situation	Use Built-in	Use Code Node
Simple data mapping	Set node	—
HTTP request	HTTP Request node	—
If/else branching	IF node	—
Multi-field transform with custom logic	—	Code node
Custom authentication signing (HMAC, AWS Sig v4)	—	Code node
Deduplicate items by a computed key	—	Code node
Parse a non-standard date format	—	Code node
Simple value extraction	Set node	—

Unlock the full lesson

You've read the first 2 sections. The rest of this lesson covers Build It, Use It, Ship It, Exercises, Key Terms, Common Pitfalls — plus a hands-on lab, quiz, and project artifact.

Create a free account to unlock Phase 0 and Phase 1 of every course — no credit card.