Secure by default
Firewall, SSH keys, TLS and updates from day one.
08 / 08 — Services
The best application is worthless on an insecure server. We set up Linux and Windows servers, harden them, automate deployments and backups and monitor operations — for web applications, databases, AI inference and telephony.
01
The typical data flow — from input to result. Every station is replaceable and adapted to your systems.
02
Linux and Windows servers, securely configured and reliably operated.
03
This is what an excerpt from a typical project looks like — readable, commented and without magic. That is exactly how we hand code over to your team.
# Basis-Härtung für einen neuen Ubuntu-Server
$ ufw default deny incoming && ufw allow 22,80,443/tcp && ufw enable
$ sed -i 's/^#\?PasswordAuthentication .*/PasswordAuthentication no/' /etc/ssh/sshd_config
$ apt install -y unattended-upgrades fail2ban && systemctl enable --now fail2ban
$ certbot --nginx -d app.example.de --redirect --hsts
$ restic backup /srv --tag nightly && restic check # Backup nur mit Restore-Test 04
Firewall, SSH keys, TLS and updates from day one.
Reproducible deployments with rollback.
Problems are reported before users notice.
Backups that are tested regularly.
05
08.1
From legacy hosting to a clean, documented server environment — with no downtime.
08.2
An in-house GPU machine that serves your local language model securely.
08.3
Open ports, outdated packages, weak configuration — found and fixed.
08.4
New versions go live at the push of a button, with rollback.
06
A chat assistant with RAG over all approved sources, per-document permission checks and answers with citations.
A GPU server with local language models, internal chat interface, user management, logging and an API for your own tools.
07
For web applications, AI inference and telephony Linux is usually first choice. Windows Server makes sense when your software or MSSQL environment requires it. We support both.
Yes, with fixed maintenance windows, monitoring and documented changes.
We check memory, graphics card and load. Existing hardware often suffices for smaller models; otherwise we recommend a targeted upgrade.
More services
LLM applications that work with your data and complete tasks on their own.
Language models on your own hardware — data-sovereign, no cloud dependency.
Business processes that run themselves — with intelligent assistants and integrations.
Ready when you are
Describe your project in a few sentences — you’ll get an honest assessment of feasibility, effort and the way forward.