SEO that brings its own software.
I combine technical SEO and GEO, AI consulting and my own software: reporting portals, AI tools for your team, automation and pages that search engines and AI agents can read. I build and run it myself.
The measurement
What machines actually read from websites.
Three numbers from my own measurements: domains tested, results counted.
- Homepages where an AI fetch failed498 of 1,000DNS errors, 403, timeouts, robots.txt. Own measurement across the Tranco top 1,000, 2026.
- Fetched homepages that returned empty text27 of 502Content loaded by JavaScript, behind bot protection or redirects. Same measurement.
- German domains with a SOVP score of 75 or more0 of 523Ten German sector lists, SOVP benchmark September 2026. Scale 0 to 100. Read the research report
In the browser the page looks fine. The fetch sees something else. What gets found is what the fetch sees.
The graph
Seven areas, one system.
Each area stands on its own. The connections between them are why I do both: the audit feeds the reporting, the reporting runs automatically, and the portal you read it in is custom software itself.
Click an area to see its capabilities.
The proof
Everything described here is already running. For my own business.
The reporting portal, the CRM and the scan engine have been running in my own operation for months. Here you see the portal that pulls Search Console every morning at 02:30 UTC.
Real values from my portal for litzki-systems.com, 90 days to 22 Oct 2026. I hide click and impression totals on purpose. The portal runs with 1,649 tests and 5 properties, daily sync.
A case from my own operation
A report that fills itself.
Every Friday, as an SEO freelancer, I pulled numbers from Search Console and GA4 into a spreadsheet. Two hours, sometimes three. If I was out, the report was out too.
A portal fetches both sources every night, computes comparisons, flags opportunities and losses, and sends a report on Monday. I read it and work with the results directly.
This is my own operation: back then as an SEO freelancer, today with my own portal.
- Check cannibalisation: 33 queries with impressions on several URLs
- Done with the relaunch: paths without trailing slash and on www now 301 to one address
- Thin pages: 61 of 125 URLs below 20 impressions
The process
Four steps. After the first you know whether the second is worth it.
1 Call free
Thirty minutes. I look at where you lose time or visibility and say openly whether I am a fit.
2 Audit fixed price
Technology, data and machine-readability of your site, as a report with a prioritised list. The result is yours.
3 Build fixed or quoted
In SEO and GEO every measure is its own module with a fixed price. Software, portals and automations I build on an individual quote. Start as small as you like.
4 Operate monthly or quoted
Monitoring, updates, adjustments. For SEO and GEO cancellable monthly, for operating custom software on an individual quote.
Cancellable monthly. Whoever stays, stays by choice.
Who it's for
For companies where visibility and software are connected.
What decides is whether your site brings revenue and whether someone in-house understands the technology, whatever your industry.
Click an audience: its areas light up.
Talk
Thirty minutes that show whether this is worth it.
I listen to where you lose time or visibility and tell you openly whether I can help. Either way, you get a clear answer.
Your call is with Thorsten Litzki.
This is how an AI reads this page.
# Litzki Systems > Technical SEO, SEO reporting, GEO, AI consulting and custom software. I build it and run it myself. ## Technical SEO Ten years in the DACH market, hands-on. - Technical SEO audit - Semantic architecture and internal linking - Structured data (JSON-LD) - Internationalisation (hreflang, canonicals) - Core Web Vitals - Migrations, sitemaps, redirects - Briefings and knowledge transfer Built with: Search Console, PageSpeed Insights, Python, JSON-LD ## SEO Reporting Search Console and GA4 in one portal, daily. - GSC+GA4 portal with 90-day dashboard - History since May 2025, longer than Search Console itself provides - URL inspection, sitemaps and Core Web Vitals on a schedule - Content decay, CTR opportunities, cannibalisation - Report API with JSON and CSV - XLSX export without SaaS - Weekly report by e-mail - Infrastructure scan module (180+ parameters) - Compliance evidence module (EU AI Act, NIS2, DORA) Built with: Python, Flask, SQLite, GSC API, GA4 API, Chart.js ## GEO & AI Visibility What ChatGPT, Perplexity and agents read from your site. - llms.txt, ai.txt, entity map and agent-readable interfaces - Identity consistency across JSON-LD, llms.txt and DNS - Rendering probe: what arrives without JavaScript - Quick check of infrastructure signals - AI crawler policy (robots.txt, ai.txt) - Benchmarks across domain lists Built with: Node, Playwright, html2text, llms.txt, JSON-LD ## Agentic SEO Data for machines that check and choose. - Measuring agentic readiness - Your offer as data: services and prices as a JSON-LD offer catalog - Key directory for bot signatures under Web Bot Auth - Disclosed bot rules: robots.txt gate, opt-out, user agent - Agent trust: contributions to Agent Manifest, SOVP bridge to TRACE - Four steps: measure, work list, implement, measure again Built with: Node, Playwright, JSON-LD, Web Bot Auth, RFC 9309 ## Custom Software Portals, dashboards and CRM, built around your workflow. - Customer portal with magic-link login - Admin dashboards with KPIs and charts - CRM with AI assistant - Checkout, Stripe webhooks, PDF delivery - Newsletter with double opt-in - PDF reports and quote generators - Signed evidence (Ed25519, OSCAL) - APIs with OpenAPI, key management, quotas Built with: Node, Express, SQLite, Stripe, Resend, Cloudflare Functions ## AI Consulting & Automation AI in daily work: skills, tools and workflows your team runs on its own afterwards. - Claude skills and workflows for your team - Your own AI tools on your data: text from the AI, numbers from measured sources - Advice on AI workflows - AI as an auditor in operations: weekly security audit of the server - AI bug hunt through the entire source code, file by file - Guardrails for AI: read-only access, no automatic model calls in customer applications - Claude API: analysis, drafts, web search with sources - Local AI tools without cloud sync - Scheduled pipelines and reminders - Domain monitoring with alerts on score drops - Retention periods as schedules: lead data after seven days, daily clean-up runs - Counter-check and learning from outcomes - macOS automation (Mail, Calendar, Contacts) Built with: Claude API, cron, systemd, AppleScript ## Web Development & Operations No framework, no WordPress, checks before every push. - Static sites on Cloudflare Pages and Workers - Security headers, CSP, caching, redirects - Automated checks before every push: canonical, hreflang, sitemap, HTML structure, scripts, data - Server operations: nginx, PM2, systemd, five services in production - Daily backups to external storage, monthly restore test - Server hardening: port scan, file integrity, SSH checks, security regression tests - Monitoring with batched alerts, only when something is found - Privacy in code: cookies, DOI, deletion - Bilingual operation DE/EN Built with: Cloudflare Pages, Workers, nginx, GitHub Actions, CSP ## Connections - seo -> reporting: Audit feeds the data - seo -> geo: GEO builds on technical SEO - reporting -> automation: Reports run on their own - reporting -> software: The portal is custom software - software -> automation: Software with agents - geo -> software: Machine-readable from day one - web -> seo: Operations carry the tech - web -> software: Operations carry the software - web -> reporting: Operations carry the portal - geo -> agentic: Agents need what AI search reads - web -> agentic: Operations deliver signatures and access rules