Insights, tutorials, and tech news
Insights, articles, and news from the tech world
20
Articles
Latest from the Blog
The server costs 5 to 20 euros a month — that is the smaller part. The full calculation including backup and operating time, and when self-hosting actually pays off.
A server in Germany does not make a Nextcloud GDPR-compliant by itself. What else is required — and which three encryption layers get confused.
Microsoft 365 wins on convenience, Nextcloud on data sovereignty and cost curve. But the decision hangs on a different question entirely — one rarely asked.
The learning platform is now an app: continue running Linux VMs on the go and validate tasks right in the app — free for iPhone and iPad, Android to follow.
AI assistants build apps in hours – and drag in security holes nobody reviews. The five most common risks and how to spot them instantly.
AWS, Stripe, OpenAI and Anthropic keys constantly end up in the browser bundle. How the leaks happen – and how to check your site for free.
A Vite or Next.js dev server instead of a production build is a critical risk. How to spot it from the outside and fix it.
Lorem ipsum leftovers, default templates, „Get started by editing“ – the typical traces of unchecked AI code and what they reveal about security.
Server full, port in use, 47 files created — Claude generates code, but cleanup is on you. The survival kit for vibe coders.
Containerization is not self-protection. We show the 6 most important measures to make Docker containers truly secure.
server_tokens off, security headers, rate limiting, block .env, modern SSL and HTTPS redirect — the 6 most important Nginx security settings with copyable configs.
SSH keys instead of passwords, firewall, Fail2ban, HTTPS, backups — the 8 most important security measures for every Linux server with copyable commands.
Feature comparison of the two AI pentest tools: community, CVE DB, GDPR default, Landlock sandbox, false-positive rate in benchmark.
Production patterns for AI agents using Anthropic Claude and FastAPI: streaming, prompt caching, tool-use validation, rate limiting, observability.
How we combine kernel-level sandboxing (Landlock LSM) with AI pentest planning (NemoClaw) — and why local LLMs are the default.
Why SaaS pentest tools aren't permitted for many German companies — and how self-hosted tooling solves it.
The methodology behind TechPuls: data sources, duplicate filtering, city assignment, and tech score calculation.
Anthropic brings Claude directly into Microsoft Word. What looks like another AI integration at first glance is actually a strategic move against Microsoft Copilot — and a turning point for everyone who works with Word daily.
Meta unveiled Muse Spark, an AI model that understands users more precisely than ever. Ad performance will rise. GDPR questions explode. We break down what this really means for German companies.
Topics
Get in Touch
Let's bring your idea to life together. We're happy to advise you with no obligation.
How do you like this page?
