senn-techWorkflow Automation with n8n — Self-Hosted
Automation without code is the premise. Zapier, Make, and Power Automate deliver this as SaaS. n8n delivers it as a Docker container — with one crucial difference: you control the data and don’t pay per step.
Nodes, Webhooks, Workflows
n8n speaks to over 400 services. Each integration is a node — Slack, email, database, REST API, Nextcloud, Odoo, you name it. Nodes are connected, with conditions in between, loops, and error handling.
The webhook node makes n8n receptive: an external service calls a URL, n8n triggers the workflow. No polling, no delay.
docker run -d --name n8n -p 5678:5678 n8nio/n8n
The deployment is one command. Workflows and credentials reside in a PostgreSQL database — the same infrastructure already running.
AI Steps: Agents in the Workflow
Since 2025, n8n has brought AI nodes: LLM retrieval, text classification, vector search, prompt chaining. A workflow can now classify incoming emails, retrieve relevant documents from Qdrant, formulate a reply via LLM, and send it back — all without code.
This is no toy. It’s a productivity leap for helpdesk automation and internal processes.
The Difference from Zapier and Make
SaaS providers bill per task. At 10,000 steps per month it’s cheap. At 500,000 it gets expensive. n8n costs nothing per step — it runs on your own hardware. The limiting factor is compute time, not the pricing model.
Add to that data sovereignty: workflow data never leaves your network. If you process customer data, you need exactly that.
Code Nodes for the Gaps
If no ready-made node exists, you write JavaScript or Python directly in the workflow. That’s the safety net that prevents a missing connector from killing the whole idea.
Conclusion
n8n is for automation what Nextcloud is for file sharing: the self-hosted alternative that matches the SaaS version in every way — except the pricing model.
Is n8n worth it over Zapier if we only have a few workflows?+
With few steps, Zapier or Make is often simpler — per-task billing stays cheap. n8n shows its strength only at higher volume: it costs nothing per step because it runs on your own hardware, and workflow data never leaves your network. If you process customer data or exceed 10,000 steps a month, n8n is the better choice.
Does our data stay in-house when self-hosting n8n?+
Yes. n8n runs as a Docker container on your own hardware, and workflows and credentials reside in a PostgreSQL database in the same infrastructure. The processed data never leaves your network — exactly the decisive point when you process customer data or must meet compliance requirements like NIS2 or GDPR. The limiting factor is compute time, not the pricing model.
What if there is no ready-made n8n node for our service?+
Then you write JavaScript or Python directly in the workflow — the safety net that prevents a missing connector from killing the whole automation idea. n8n speaks to over 400 services natively, and the webhook node also makes it receptive to arbitrary external calls. The gap can therefore always be closed productively.