4.1 · n8n: open source & self-hosting

Level 4 · Expert: n8n & AI in workflows

4.1n8n: open source & self-hosting

Goal: Understand n8n's positioning (open source, self-hostable, sovereign) and when to prefer it over the Make/Zapier SaaS platforms.
Estimated time: 12 min · self-paced reading · If you only have 5 min: read “The essentials in 30s” + the diagram.

The essentials in 30 seconds

  • n8n = the open-source alternative to Make/Zapier: you can host it yourself, with no per-operation billing.
  • Self-hosting = sovereignty + controlled costs: total control over data, but a server to administer (updates, backups, security).
  • Lina's reflex: prototype in the cloud to learn, switch to self-hosting only when sensitive data or volumes justify it.

Running case — Lina compares options for her company in Oran

Lina, digitalisation officer in an SME in Oran, already automates with a SaaS platform: the bill climbs with volumes and management worries about HR files passing through foreign servers. She tries n8n in its cloud version on a small workflow (trigger → action), recognises the node logic, then quantifies the gain: high volumes + sensitive data + future AI calls. She proposes prototyping in the cloud and moving to self-hosting only with the IT department's approval. First milestone of the thread: choosing the right tool before plugging in AI.

This same case returns in the lesson, the diagram, the script and the 48h challenge — and continues in 4.2, 4.3, 4.4 and 4.5.

1. Sovereignty and controlled costs

n8n is the open-source alternative to Make and Zapier. Major difference: you can host it yourself (self-hosting) on your own server, giving you total control over data and removing per-operation billing. For sensitive data (health, HR, finance), high volumes or deep customisation, it is often the best option. Like Lina, compare the SaaS bill against data criticality before deciding. A managed cloud version also exists to get started without a server.

2. Node logic, technical orientation

n8n follows the same visual logic (nodes connected into a workflow) with a more “technical” orientation: easy access to HTTP calls, to code (JavaScript/Python) when needed, and an active community. Lina finds her Make reflexes there, with extra doors into APIs.

It is the tool of choice for AI workflows: it flexibly handles model calls, data and secrets (see 4.2 and 4.5). In short: power, sovereignty and controlled costs, at the price of a little more technicality — especially when self-hosting.

3. Prototype in the cloud, self-host if justified

Self-hosting is not compulsory: it requires skills (or a provider) for server updates, backups and security. Lina's practical rule: start in the cloud to learn and prototype, move to self-hosting only when data sovereignty or large-scale costs truly justify it — with the IT department's approval.

Common misinterpretation: «Open source means free and constraint-free: I install n8n on an old PC and that's it». No. The software itself is open, but self-hosting has a real cost: a server to administer, updates, backups, security to maintain — otherwise “sovereignty” becomes a sieve. Open source = independence and transparency, not zero effort.
Why n8n? The sovereign alternative Open sourceSelf-hostable, no per-operation billing+SovereigntyTotal control over data+AI-orientedHTTP, code, secrets Start in the cloud to learn; self-host when justified.
The sovereign alternative
Text version of the diagram (SEO / screen readers / translation): 1) Lina compares: her SaaS bills by volume and routes sensitive HR files externally. 2) She tries n8n in the cloud and recognises the node logic, plus HTTP and code. 3) She only self-hosts if justified (sensitive data, high volumes), with IT approval. Lesson: power and sovereignty at the price of a little more technicality.

🔎 Concrete example — 🔎 Concrete example — When to choose n8n

Lina's grid, filled in with her real case in Oran:

Sensitive data (HR, payroll) → self-hosting (sovereignty)
High volumes (SaaS too expensive) → no per-operation cost
AI workflows, APIs, custom code → technical orientation
→ otherwise, Make/Zapier remain simpler to start with.

Reading: three situations where n8n wins, one where SaaS is enough.

Lesson: prefer n8n when sovereignty or volume come first.

Section vocabulary

n8n
Open-source, self-hostable, node-based automation tool; the sovereign alternative to Make/Zapier.

Simple example: Lina recreates her trigger → action workflow in it to compare ergonomics.

What it is not: a plain SaaS connector with no hosting option.

Reflex: test the same small workflow in the cloud before deciding.

Open source
Software whose code is open and freely usable; here, a guarantee of transparency and independence.

Simple example: Lina can audit what the tool does instead of taking it on trust.

What it is not: totally free of charge with no hosting or maintenance costs.

Reflex: distinguish open licence from operating cost.

Self-hosting
Hosting the software on your own server to control data, costs and customisation.

Simple example: Lina's SME's HR files stay on the company's server.

What it is not: using the managed cloud version without administering a server.

Reflex: only self-host with skills or a provider + IT approval.

Node
A building block of an n8n workflow performing one operation: trigger, action, HTTP call, transformation, code.

Simple example: Lina's workflow: form trigger node → database write node.

What it is not: a complete server or an entire piece of software.

Reflex: read a workflow node by node, left to right.

Data sovereignty
Total control over where and how data is stored and processed; key for sensitive data.

Simple example: Lina keeps HR data under control instead of seeing it transit abroad.

What it is not: encryption alone, without control of hosting.

Reflex: ask for each workflow: where does the data really go?

In practice — In practice — Discovering n8n and its benefits

  1. Try n8n in its cloud version (or a demo instance) to rediscover the node logic seen with Make.
  2. Recreate a simple workflow (trigger → action), like Lina, to compare ergonomics.
  3. Identify a case where n8n wins for you: sensitive data, high volumes, or a need for API calls and custom code.
  4. Note the self-hosting option: total control over data and no per-operation billing, but a server to administer — to validate with your IT department.
You understand n8n's positioning (open source, self-hostable, sovereign) and when to prefer it over the Make/Zapier SaaS platforms.

Ready-to-say script — proposing n8n to IT (30 s)

«This is Lina: our volumes are driving up the SaaS bill and our HR files transit through external servers. I propose prototyping our workflows in n8n in the cloud, then studying self-hosting with you if the test confirms the gain — updates, backups and security included in your decision.»

Adapt to your context; promise neither zero cost nor deadlines, ask for the technical decision.

Copy-paste template — n8n vs SaaS choice grid (fictitious example)

CHOICE GRID — fictitious example to adapt
NEED: _______________
SENSITIVE DATA (HR / health / finance)? yes / no
VOLUMES (operations / month, order of magnitude): _______________
API CALLS / CUSTOM CODE? yes / no
→ IF sensitive OR high volumes OR AI+API: test n8n (cloud then self-host)
→ OTHERWISE: stay on the simple SaaS (Make / Zapier)
IT APPROVAL REQUESTED ON: _______________

Very small business with no server or IT department

Stay on managed cloud (n8n or SaaS): self-hosting without skills or a provider creates more risks than it removes.

SME with sensitive HR files

Like Lina in Oran: prototype in the cloud with fictitious data, then self-host with IT for real data.

High volumes, exploding SaaS bill

Cost the per-operation price over 12 months: that calculation alone often justifies moving to n8n.

Junior team, simple need

Don't impose n8n on principle: if SaaS is enough, its simplicity beats unused power.

Caution: Self-hosting means administering a server (updates, backups, security, monitoring): without skills or a provider, do not self-host real data. As soon as sensitive data transits (HR, health, finance), have the architecture validated by your DSI/RSSI and stay generic about hosting. In Algeria, processing personal data falls under Law 18-07: check with your DSI/RSSI or a local legal expert where your data may legally reside. Educational content, not legal advice nor an installation guide.

✓ Best practices

  • Prototype in the cloud, self-host if justified.
  • Prefer n8n for sensitive data and high volumes.
  • Use its technical orientation for AI.

✗ Common pitfalls

  • Self-hosting without skills or a provider.
  • Neglecting updates, backups, security.
  • Choosing n8n for a simple need.

Key takeaways

  • n8n is the open-source, self-hostable alternative to Make/Zapier.
  • Self-hosting gives total control over data and removes per-operation billing.
  • Ideal for sensitive data, high volumes and deep customisation.
  • Node logic, technical orientation (HTTP, code), excellent for AI workflows.

48h challenge — apply tomorrow

Recreate 1 small real workflow in n8n in the cloud (trigger → action) and fill in the choice grid with your case: is n8n justified for you?

Section exit checklist
  • I can explain open source, self-hosting and sovereignty.
  • I can cite 3 cases where n8n wins.
  • I have tested 1 workflow and filled in the choice grid.
Self-rating 1-5: do I know when to prefer n8n over SaaS?have I identified the technical approval needed for self-hosting?

Frequently asked questions

Is self-hosting n8n only for technicians?

Installing and maintaining a server requires minimum skills (or a provider): updates, backups, security. Guides and images exist, but they cost time. Good news: n8n also offers a managed cloud version. Practical rule: start in the cloud to learn; move to self-hosting only when sovereignty or large-scale costs justify it.

Why is n8n often recommended for AI?

Because it flexibly handles what AI workflows need: HTTP calls to models, fine-grained data handling, secret management and no per-operation billing. Add data sovereignty — precious when sending internal documents to an AI — and n8n ticks almost every box for AI automation (see 4.2 to 4.5).

Check your understanding

What is n8n's major distinctive feature?

Additional resources (optional, to go further)

Level 4 examFresh, harder questions to test what you've learned