AI & GenAI
Artificial Intelligence (AI) covers the practical application of Large Language Models (LLMs), RAG architectures, and autonomous AI agents to automate business processes and work efficiently with data.
Kaizen and Agentic Development
Kaizen means continuous improvement: take one small step forward every day. Don’t wait for ideal conditions or a big breakthrough — improve something by one percent today. This principle turns out to fit today’s agentic development and vibe coding surprisingly well — the style of work where a developer describes tasks in plain language and an AI agent writes, tests, and adjusts the code itself. 1% every day Faced with a project we want to build, we can tell ourselves:
Critical External Dependency: When a Vendor Becomes a Risk
Every company today depends on external technologies and vendors. We use cloud, SaaS services, payment gateways, accounting systems, hosting, or external developers. That is not a problem in itself. The problem starts the moment we cannot function without one particular vendor. That is called a critical external dependency. When does it become a problem? Picture an e-shop that has been running on a single platform for several years. It holds the website, the products, the orders, the stock, and the link to accounting.
AI Helpers for Content Management
Status: Production The client has a content management system for multiple websites. Article authors previously had to manually fill in the SEO title and description, make corrections, search for synonyms, translate text, and perform other text-related tasks. I proposed and developed AI assistants that perform specific operations with a single click of a button. Key features One-click AI-powered SEO autofill: titles, descriptions, and metadata Grammar checking and text corrections Synonym suggestions and text improvements Automated translation between languages Custom prompt-defined text operations via configurable buttons Helped authors write articles faster and with higher quality
SEO tools
Status: Under active development On-page SEO analyzer and site auditor. Crawls websites to surface metadata, content, and technical SEO issues, with AI/GenAI content readiness checks. Key features Crawlee + Playwright crawler extracting SEO meta tags, structured data (JSON-LD/microdata), and AI-indexing metadata Reports: Markdown SEO audit, title/meta-description issue lists, 404 and broken-link checks, empty-href scan Optional LLM-powered title/meta-description fix suggestions (OpenAI-compatible, incl. local models) MCP server exposing crawl and report tools to AI agents, plus an SEO-consultant persona Smart incremental crawling, sitemap discovery, and configurable rate limiting Dockerized stack; English and Czech report output Open source. The code is on GitHub and the running tool is at seo.ludekkvapil.cz — free to use, and a working sample of how I build.
OSINT tools
Status: Under active development AI-orchestrated OSINT investigation platform. Plans and executes multi-step intelligence cases — collecting evidence, generating reports, and maintaining a persistent knowledge wiki. Key features LLM-driven case orchestrator: auto-plans queries, runs tools, writes structured reports 15+ lookup scripts: RDAP, WHOIS, Shodan, ARES, LinkedIn/Apify, VIN, username recon, web fingerprinting, Czech public registries, media monitoring MCP-exposed tools: plan, investigate, run single tool, case status, wiki lookup/ingest Persistent LLM wiki for cross-case knowledge retention Optional Kali MCP pentest server for authorized active recon Dockerized stack with OpenClaw gateway; works locally or with cloud LLMs Open source. The code is on GitHub .
ClawBot – Agentic Ops Manager (IronClaw)
Status: Active development ClawBot is an agentic Ops Manager built on IronClaw. It orchestrates a team of specialized agents behind a unified MCP gateway, handling weekly SEO audits, daily OSINT monitoring, real-time support, and automatic knowledge base updates. It also manages the onboarding of new e-commerce stores and keeps Drupal and Hugo content in sync. Key features Unified MCP gateway with 20+ tools (RAG CMS, SEO crawler, OSINT, Sales) Scheduled tasks: weekly SEO audits, daily OSINT monitoring, and daily support gap analysis CMS/RAG integration: Drupal content synchronized with the knowledge base Automation: background tasks, reminders, and Canvas previews Docker-based stack for both local development and CI
GenAI, Chatbots & Agents
I help companies deploy generative AI into real processes—from chatbots over internal documentation to autonomous workflow agents and LLM integrations in existing applications. I build solutions that are secure, measurable, and maintainable over time. This page is the technical one. What AI does on your website and inside your company, without the jargon, is on AI Assistant for Your Website and Documents . How I build applications and integrations, AI included, is on the Agentic Development of Applications and Integrations page.
AI Assistant for Your Website and Documents
Most companies wrote down the answers to their common questions long ago — on the website, in manuals, in the price list, in internal policies. Nobody finds them there. A customer asks support something the third paragraph of a guide already answers, and a new colleague asks about what sits in a shared folder. I integrate AI into your Drupal website so that it answers from your own texts. The assistant knows only what you give it, and every answer says where it came from.
Vibe-Coded App Review
You built your app with Claude Code, Cursor, Lovable, Bolt or another AI agent, and it works. Before you let it near your customers and their data, though, it is worth knowing what is really inside. AI writes fast and convincingly, but it can just as quickly skip a permission check or put a key straight into the code. I use vibe coding every day and teach it, so I know where agents go wrong most often.
A Knowledge Base for AI in Drupal: RAG and MCP in Practice
A Drupal knowledge base is where a company maintains, approves, and keeps its knowledge current. An AI assistant built on it answers from content somebody stands behind, rather than from a pile of documents. In Which Open-Source CMS Should You Choose for an AI Knowledge Base? I compare five systems. This article goes deeper into one of them: how to design a knowledge base in Drupal and connect it to AI through an API, RAG, and MCP.
Which Open-Source CMS Should You Choose for an AI Knowledge Base?
Most companies introducing an AI assistant over internal knowledge start with technology: which model, framework, or vector database to use. Those layers can be changed when needed. A knowledge base and CMS give an AI project a stable foundation: they determine where company knowledge is created, who maintains it, and how it stays current and approved. AI then works with information the company can trust. A CMS as a knowledge base, not only a website In an AI project, a CMS has a different role than on a website. It is where knowledge is created, structured, approved, and maintained, and where AI retrieves it. The selection criteria therefore differ: