Dygistic Labs · All work

Building my own SEO automation stack with n8n

Why I run automation on my own server instead of stacking SaaS subscriptions, how the stack is set up, and what it runs today.

Period
April 2025 to present
Project
Dygistic (own)
Stack
VPS, Docker, n8n, Nginx, Let’s Encrypt
Disciplines
Automation, integrations, AI agents
In short

Own the plumbing, rent nothing you do not need.

SEO work is full of repeatable steps: pulling data, checking pages, formatting reports, publishing content. I built a self-hosted automation environment so those steps run without me, and without a per-task SaaS bill.

27
active workflows (34 built)
12
MCP servers for AI assistants
15 min
between uptime and noindex checks

The problem with renting every step

Most SEO tooling is excellent at one thing and charges for it monthly. Chain five of them together and you have a fragile, expensive process that still needs a human to copy data from A to B. I wanted a single place where workflows live, talk to any API, and can call an AI model when judgement is needed.

What I set up

  • An Ubuntu VPS with Docker Compose running n8n on a persistent volume, so upgrades never lose workflow data.
  • Nginx as reverse proxy on a dedicated subdomain, HTTPS via Let’s Encrypt, HSTS and authentication in front of the editor.
  • OAuth connections to Google, WordPress via the REST API with application passwords, and database access to client sites where the client agreed to it.
  • Guardrails by default: read-only unless a write is needed, write actions limited to a whitelist, destructive and financial actions blocked, and my approval before anything is published or booked.

What runs on it

The stack powers the operational side of Dygistic. Monitoring: every 15 minutes each active client site is checked for uptime and an accidental noindex, and I only get an email when something changes. For one large webshop it also compares key pages, robots.txt and sitemaps with the day before every morning and flags URLs that disappeared without a redirect. Reporting: a work log that every working session writes to, turned into a Friday overview with work per client, draft client updates and the status of all monitored sites. Integrations: MCP servers that let an AI assistant work on WordPress sites, databases, hosting and DNS, bookkeeping, Search Console and GA4, each with its own limits. Data: weekly Search Console snapshots per tracked URL, written to a Google Sheet with a history line per week.

The free SEO tools on this site and the glossary tooltips came out of the same lab mindset: build the thing once, reuse it everywhere.

Before and after

Before
Every recurring task was manual or spread over separate SaaS tools with their own logins, limits and invoices.
After
One self-hosted environment runs the recurring work, connects to any API, and hands me a draft or an alert instead of a to-do.

What this means for you

If your marketing team copies data between tools every week, that process can almost certainly be automated. I build these pipelines for clients as part of my marketing automation work, from a single reporting flow to a full content pipeline.

Tired of copy-pasting between dashboards?

Tell me which weekly task annoys you most. I will tell you whether it can be automated, what it would take and what it saves.

Start a conversation
One consultant, start to finish.
Developer-ready deliverables.
NL, BE, DE and UK markets.
Scroll to Top