Short answer: One system holds the real stock number and every sales channel reads from it. A sale on any platform fires a webhook, the hub decrements the central count, and pushes the new figure to the other three within seconds. Product content is written once and adapted per marketplace. Built on n8n with Airtable as the source of truth.

What this is: a reference build. The architecture, the conflict-resolution rules and the failure handling, written up so you can judge the engineering. There are no client figures here. What you can check yourself is the running system: try the demos →.

The problem it solves

Selling on Shopify, WooCommerce, Amazon and eBay means four systems that each believe they know your stock level, and none of them talk. Sync happens when somebody remembers to do it, so the true number lives in a spreadsheet that is already out of date. The result is overselling, worst on the days that matter most, and a product description that has quietly drifted apart across four listings.

The fix is not faster syncing. It is deciding that exactly one system holds the truth and that every other system reads from it. Everything below is the consequence of that one decision, including the awkward parts: what happens when two platforms sell the last unit in the same second, and what happens when a marketplace API is down.

At a glance


Stack	Central n8n hub + Airtable source of truth + real-time webhooks + AI listing content
Sync direction	Platforms report sales in; the hub pushes authoritative stock out
Guardrails	Optimistic locking on the central count, per-platform retry queue, buffer stock during peaks, reconciliation sweep
You can inspect	The full n8n JSON, exported from the running instance

This is the kind of build I do under integrations & APIs, connecting tools so data moves on its own. See the live demos for working examples.

The Solution

A central inventory hub with real-time sync and AI-assisted product management.

Tool Stack

Component	Tool	Why
Central Hub	Airtable / Notion	Flexible, API-friendly, visual
Workflow Engine	n8n	Real-time webhooks, self-hosted
Shopify Sync	n8n + Shopify API	Native integration
WooCommerce Sync	n8n + WooCommerce API	REST API
Amazon Sync	n8n + Amazon SP-API	Marketplace integration
AI Product Assistant	Claude API	Categorization, descriptions
AI Local	Ollama	Anomaly detection, cost savings

Architecture Overview

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│   Shopify   │◄───►│             │◄───►│ WooCommerce │
└─────────────┘     │   Central   │     └─────────────┘
                    │   n8n Hub   │
┌─────────────┐     │             │     ┌─────────────┐
│   Amazon    │◄───►│             │◄───►│    eBay     │
└─────────────┘     └─────────────┘     └─────────────┘
                          │
                    ┌─────┴─────┐
                    │   Airtable │
                    │  (Source)  │
                    └───────────┘

Core Features

1. Real-Time Inventory Sync

How it works:

Each platform sends webhooks on order/inventory change
n8n receives event, updates central Airtable
Change propagates to all other platforms within 60 seconds

Webhook Events Handled:

Order placed → Decrement stock across all platforms
Order cancelled → Increment stock
Manual adjustment → Sync to all platforms
Refund processed → Restock if applicable

Conflict Resolution: When two orders come in simultaneously for the last item:

First webhook wins (atomic update in Airtable)
Second order triggers stock alert
Human decides: fulfill from alternate source or cancel

2. AI-Powered Product Categorization

New products need proper categorization for each marketplace. AI handles this automatically.

Input (from supplier data):

Product: "Ergonomic Wireless Mouse, 2.4GHz, Silent Click, Black"
EAN: 1234567890123
Brand: TechFlow

AI Output:

{
  "amazon_category": "Electronics > Computers > Mice",
  "amazon_keywords": ["wireless mouse", "ergonomic mouse", "silent mouse"],
  "shopify_type": "Computer Accessories",
  "shopify_tags": ["wireless", "ergonomic", "mouse", "office"],
  "ebay_category_id": 23160,
  "attributes": {
    "connectivity": "2.4GHz Wireless",
    "color": "Black",
    "feature": "Silent Click"
  }
}

The AI maps products to each platform’s taxonomy automatically.

3. Stock Anomaly Detection

Ollama runs locally to detect inventory anomalies:

Monitored Patterns:

Sudden large stock decreases (theft? bulk order?)
Stock count doesn’t match sales velocity
Platform counts diverge from central hub
Unusual refund patterns

Alert Example:

⚠️ Anomaly Detected
Product: SKU-1234 (Wireless Mouse Black)
Issue: Shopify shows 45 units, central hub shows 23
Action: Manual stock check required
Possible cause: Unrecorded shipment or sync failure

4. AI Description Generation

Creating marketplace-specific listings used to take 30 minutes. Now:

Input: Basic product info + one photo

AI Generates:

Amazon listing (bullet points, A+ content ready)
Shopify description (brand voice, SEO optimized)
eBay listing (marketplace-specific keywords)
Translations (German, French, Spanish)

Before → after:

Input:

Mouse, wireless, ergonomic, black, silent clicks

Shopify Output:

Work in comfort with the TechFlow Ergonomic Wireless Mouse.
Designed for all-day use, featuring silent click technology
that won't disturb your colleagues. The 2.4GHz wireless
connection provides reliable, lag-free performance up to 10m.

✓ Ergonomic design reduces wrist strain
✓ Silent switches for quiet operation
✓ 12-month battery life
✓ Plug & play USB receiver

Amazon Output:

• ERGONOMIC COMFORT: Scientifically designed shape supports
  natural hand position, reducing fatigue during extended use
• SILENT CLICK TECHNOLOGY: Noise-reduced switches perfect
  for office, library, or shared workspaces
• RELIABLE 2.4GHz WIRELESS: Strong connection up to 33ft,
  no interference from other devices
...

Implementation Highlights

Shopify Webhook Setup

// n8n webhook node configuration
{
  "topic": "orders/create",
  "address": "https://n8n.yourdomain.com/webhook/shopify-orders",
  "format": "json"
}

When an order arrives:

Extract line items and quantities
Update Airtable stock counts
Trigger sync to WooCommerce, Amazon, eBay
Log transaction for audit trail

Amazon SP-API Integration

Amazon requires OAuth and signed requests. n8n handles:

Token refresh (automatic)
Request signing
Rate limiting (throttle to avoid 429s)
Feed submission for bulk updates

Handling Platform Delays

Not all platforms update instantly:

Shopify: Real-time webhooks ✓
WooCommerce: Real-time webhooks ✓
Amazon: Feeds processed every 15 min
eBay: API calls, near real-time

For Amazon’s delay, we maintain a “pending” state:

Immediately mark stock as “reserved” in central hub
Confirm when Amazon feed completes
Alert if feed fails

What changes, and what does not

I am not going to give you a before-and-after table. I have not run this hub on your catalogue, and numbers invented for a fictional retailer would tell you nothing.

What the architecture changes is structural:

There is one stock number. Four systems stop each holding an opinion, and drift stops being possible by construction rather than by discipline.
Sync stops waiting for a person. A sale propagates on a webhook in seconds, so the exposure window shrinks from a day to the length of an API call.
A new channel is a connector, not a rewrite. Adding a fifth marketplace means teaching the hub one more output, not another spreadsheet column.

What it does not do is make overselling impossible. Two platforms can still sell the last unit in the same second, and a marketplace API can still be down when you need to push. That is what the buffer strategy and the retry queue below are for, and they reduce the window rather than close it. Anyone who tells you a sync hub gets you to zero oversells has not run one on Black Friday.

The number worth measuring is your current discrepancy rate: count how far the spreadsheet is from reality on a random Tuesday, then again on your busiest day. The gap between those two is the size of the problem you are actually buying a fix for.

Technical Details

Stock Buffer Strategy

To prevent overselling during sync delays:

Set platform stock = central stock - buffer
Buffer varies by velocity: fast sellers get larger buffer
Auto-adjusted weekly based on sales patterns

Multi-Language Product Content

For German and French markets:

AI generates base description in English
Claude translates with marketplace-specific terminology
Human review for top 20% sellers, auto-publish rest

Monitoring Dashboard

Real-time visibility in Airtable:

Stock levels across all platforms
Sync status (last update, any failures)
Anomaly alerts
Sales velocity per channel

Costs

Item	Monthly
Airtable Pro	€20
n8n self-hosted	€0
Claude API (descriptions, categorization)	€45
Ollama (anomaly detection)	€0
Total	€65/month

vs. 15 hours/week of manual work = €1,500/month equivalent.

Key Learnings

Webhooks > Polling: Real-time sync prevents most oversells
Single source of truth: Airtable is authoritative, platforms are mirrors
Buffer for slow platforms: Amazon’s 15-min delay needs accounting
AI for tedious, not critical: Use AI for descriptions, not stock counts

Build This Yourself

Here’s how to build a multi-platform inventory sync from scratch.

Node-by-Node Breakdown

1. Universal Webhook Intake

A single webhook endpoint receives events from all platforms. The first step normalizes different payload formats:

// Detect platform from payload structure
if (body.topic?.includes('shopify')) platform = 'shopify';
else if (body.source === 'woocommerce') platform = 'woocommerce';
else if (body.NotificationType) platform = 'amazon';

Output: Consistent { sku, quantity_change, new_quantity, source } regardless of origin.

2. Central Hub Lookup (Airtable)

Query your central inventory database by SKU. Airtable (or Notion) serves as the single source of truth:

Current stock level
Platform-specific IDs (Shopify variant ID, WooCommerce product ID, Amazon ASIN)
Last sync timestamp
Buffer quantities per platform

3. Stock Calculation

Handle two update types:

Absolute: “Set stock to 50” → new_quantity = 50
Relative: “Order placed, -1” → new_quantity = current - 1

Always enforce non-negative values. Log the delta for audit trails.

4. Update Central Hub

Write the new stock level back to Airtable with:

Updated quantity
Timestamp
Source platform (to prevent echo loops)

5. Platform Distribution (Split → Switch)

For each platform that needs updating (all except the source):

Shopify: GraphQL API or REST /variants/{id}/inventory_levels
WooCommerce: REST API PUT /products/{id}
Amazon: SP-API inventory feed (batch, 15-min processing)

Handle platform-specific quirks:

Shopify needs location_id for multi-location inventory
Amazon requires XML feed format and polling for confirmation
WooCommerce stock management must be enabled per product

6. Confirmation & Alerting

After successful sync:

Slack notification with SKU, new level, platforms updated
Webhook response confirms completion
Log for debugging sync issues

AI-Powered Categorization (Optional Node)

When listing new products, Claude can auto-categorize:

Amazon category path and keywords
Shopify product type and tags
Attribute extraction from product name

This node is disabled by default in the starter, enable it for new product workflows.

Get the Starter Workflow

Not a screenshot: the real workflow. This is importable n8n JSON: read every node, wire in your own credentials, run it. It is the build as designed rather than an export from a running instance, so treat it as a starting point you can verify by reading, not as a system with production mileage on it. The workflows behind the live demos are the ones exported from my own n8n.

Download n8n-ecommerce-sync.json

Quick Setup:

Import JSON via n8n Settings → Import Workflow
Configure credentials (Airtable, Shopify, WooCommerce, Slack)
Set up Airtable base with: SKU, stock, shopify_id, wc_id, amazon_asin
Configure webhooks in each platform to point to your n8n endpoint
Test with manual stock adjustments

This starter handles the core sync loop. A production system would add stock buffers for slow platforms, conflict resolution for simultaneous orders, anomaly detection (Ollama), multi-location support, and retry logic for API failures, the resilience layer that handles Black Friday traffic without breaking a sweat.

Your turn

Selling across multiple platforms?

Map the mess: Which platforms? What’s the current sync method?
Find the pain: Oversells, listing time, discrepancies?
Start with sync: Fix inventory first, then add AI features.

If keeping systems in step is a recurring headache, that is integrations & APIs work.

Book a free strategy call, I’ll assess your multi-platform setup and recommend a sync strategy.

Reference Build: Multi-Platform Inventory Sync for Retail