Shafail Shafa
Home / Services / Automation & private AI
◈ AUTOMATION & AI

Self-hosted n8n and local LLMs, and the server work underneath them.

I deploy self-hosted automation and private AI on servers you control, then keep them running. n8n with Docker, PostgreSQL, SSL and verified webhooks. Ollama and OpenWebUI for local models, with cloud APIs wired in only where you decide you want them. I run this stack in production myself, and the reliability work underneath it is the same Linux administration I have done for thirteen years.

DELIVERABLES

Self-hosted n8n

n8n on Docker behind a reverse proxy, with SSL, PostgreSQL and HMAC-verified webhooks. Set up to run in production rather than as a demo install, which mostly means the parts nobody sees: restarts, log rotation, and credentials that are not sitting in plain text.

Private AI stack

Ollama and OpenWebUI running local models on your own hardware, with cloud APIs wired in only where you decide you want them. Useful when the data involved is the reason you cannot use a hosted service.

Backups, monitoring and updates

Automated backups with the restore actually tested, monitoring with alerts, and updates applied rather than deferred. This is the part that decides whether the stack is still working in six months.

ALWAYS INCLUDED
firewall, SSL and locked-down SSH backups with a tested restore monitoring with alerts 3 days of support after delivery

HOW IT WORKS

1

Scope

We work out what gets automated, what data is involved, and what must never leave your server. Before anything is installed you get a plan and a fixed definition of what finished means.

2

Deploy

The stack goes onto a hardened server: Docker, reverse proxy, SSL, database, firewall. Credentials are stored properly, nothing is exposed that does not need to be, and nothing runs as root that does not need to.

3

Prove

I restore the backups onto a clean server to confirm they work, and I trigger real alerts to confirm the monitoring reaches me. You see the system recover before you start relying on it.

4

Hand over

A full handover, with documentation covering what runs where, how to restart it and how to restore it. If you would rather not hold that yourself, I can stay on for updates, monitoring and changes to the workflows.

Production self-hosted n8n with tested disaster recovery
An n8n instance running in production on my own hardened VPS, with a restore I've tested end to end on a clean server.

WHAT PEOPLE SEARCH FOR

n8n automation
Workflows running on a server you own instead of a monthly seat, with the data staying there too.
self hosted n8n
The community edition installed properly: Docker, a database that is not SQLite, SSL, and backups that have been restored at least once.
n8n vps setup
The server underneath the workflows. Firewall, reverse proxy, certificates, and the restarts nobody thinks about until the first reboot.
local llm server
Models running on hardware you control, for teams whose data is the reason a hosted service is not an option.
ollama setup
Ollama and OpenWebUI installed on a Linux box, reachable over SSL, with sensible limits on who can use it.
zapier alternative
The same automations moved off a per-task subscription and onto a server with a fixed monthly cost.
docker deployment
Services in containers behind a reverse proxy, with SSL that renews itself and logs that do not fill the disk.
api integration
Two systems that were never meant to talk to each other, made to, and kept working after the first API change.

Tell me what you want automated.

What the manual process is now, what data it touches, and whether it has to stay on your own hardware.

Get in touch →