# flize

> flize is GEO software for companies and agencies. It scouts the real questions and prompt trends of a target audience (Google, Brave, Reddit, gutefrage.net, LinkedIn, X, YouTube and curated trade media per industry), matches them against the website (which section of which page answers what: answered / partial / gap) and automatically produces ready-to-publish, source-backed articles for every gap. Includes a GEO readiness score (structure, content coverage, machine readability), a recurring digest and a chat integration over MCP (flize in chat) that exposes the same project data and actions to Claude, ChatGPT and other MCP-capable assistants. White-label and multi-client for agencies. Made in Germany, a product of sitebrunch GmbH.

## Full text

- [Alle Wissen-Artikel als Volltext (Markdown)](https://flize.io/llms-full.txt): Complete text of every article in the Wissen section, in one Markdown file, so an assistant does not have to crawl the individual pages.

## Pages

- [flize Startseite (DE)](https://flize.io/): German homepage: question scouting, matching, the four building blocks (GEO score, scouting, answer check, content production), for companies & agencies, pricing, FAQ.
- [flize Home (EN)](https://flize.io/index-en.html): English homepage: same content as the German page.
- [Preise (DE)](https://flize.io/pricing.html): Pricing: subscription plans with monthly credits. Free: €0, 1 project, 50 URLs, 1 free article, one-shot question research. Starter: €29/month, 1 project, 25 credits/month, 1,000 URLs, weekly trends & alerts, flize in chat (MCP). Pro: €79/month, 3 projects, 100 credits/month, 4,000 URLs. Actions such as content production or extra topic scouting consume credits, more or fewer depending on scope. Add-ons: extra project €12/month (incl. 1,000 URLs); extra credit packs +10 for €8, +50 for €35, +100 for €69, +250 for €169 per month.
- [Pricing (EN)](https://flize.io/pricing-en.html): English pricing page (same plan model).
- [Euer nächster Kunde ist kein Mensch (DE)](https://flize.io/kein-mensch.html): Explainer on the shift to the agentic web: how AI agents read today, act via Web-MCP next, and complete tasks end-to-end in future; with infographics.
- [Your next customer isn't human (EN)](https://flize.io/kein-mensch-en.html): English version of the agentic-web explainer (read → act → complete, Web-MCP, infographics).
- [flize im Chat (DE)](https://flize.io/mcp.html): How flize is operated from an AI assistant over MCP (Model Context Protocol): ask for this week's scouted topics already matched against the site, open questions with their status, the GEO score and its top fixes; brief and generate an article in conversation; run market research across every feed flize follows (Pro only); read structural fixes as before/after, apply them and mark them done. Asking costs nothing, credits are only spent on actions that cost credits in the app. Included from the Starter plan. Works with Claude, ChatGPT and any other assistant that supports MCP.
- [flize in chat (EN)](https://flize.io/mcp-en.html): English version of the MCP page.
- [Artikel schreiben (DE)](https://flize.io/artikel-schreiben.html): How an article is briefed at flize, in five steps: starting point (a suggestion from the project, an own theme, a freely described request, or an open question from scouting), angle (a real situation people come to a site for, not a keyword), goal (get cited / sell something specific / build authority), article type (answering a question needs an external source for every claim about the world; an ideas list leans on your own pages), and optional details (style guide for your voice, target market of country, language and region, a focus, and up to three of your own sources, where pasted text counts as your own words and linked pages are read and quoted like any other source). Output: direct answer first, source and date per section, FAQ plus schema markup. Content production consumes credits, more or fewer depending on scope; the final review stays with the customer.
- [Writing articles (EN)](https://flize.io/artikel-schreiben-en.html): English version of the article briefing page.
- [GEO-Tools im Vergleich (DE)](https://flize.io/geo-tools-vergleich.html): Fair comparison with AI-visibility monitoring tools (Peec AI, Otterly.AI, Rankscale): flize covers the pre-prompt phase (question scouting, answer check, content production), monitoring tools cover measurement afterwards; ideally combined.
- [GEO tools compared (EN)](https://flize.io/geo-tools-vergleich-en.html): English version of the comparison page.
- [Über uns (DE)](https://flize.io/ueber-uns.html): About the team behind flize: sitebrunch GmbH (Hamburg), specialists in digital accessibility (WCAG monitoring, EN 301 549/BFSG audits), UX and website analytics; why GEO completes the picture.
- [About us (EN)](https://flize.io/ueber-uns-en.html): English version.

## Wissen

- [Wissen (DE)](https://flize.io/wissen/): Articles on GEO, AI search, Web-MCP and the agentic web. AI-ready format (direct answer, TL;DR, sources).
- [Resources (EN)](https://flize.io/wissen/index-en.html): English knowledge index.
- [Was ist GEO? (DE)](https://flize.io/wissen/was-ist-geo.html): What GEO (Generative Engine Optimization) is, how it differs from SEO, how to start. Sources: Imperva 2025, Gartner 2025.
- [What is GEO? (EN)](https://flize.io/wissen/was-ist-geo-en.html): English version.
- [GEO vs. SEO (DE)](https://flize.io/wissen/geo-vs-seo.html): Rankings vs. citations: what stays the same, what GEO adds (answer blocks, dated evidence, machine readability) and how to prioritize both.
- [GEO vs. SEO (EN)](https://flize.io/wissen/geo-vs-seo-en.html): English version.
- [Was ist llms.txt? (DE)](https://flize.io/wissen/was-ist-llms-txt.html): What llms.txt is (curated Markdown overview for LLMs, Answer.AI proposal 2024), how it differs from robots.txt and sitemap.xml, and how to create one.
- [What is llms.txt? (EN)](https://flize.io/wissen/was-ist-llms-txt-en.html): English version.
- [Wer liest deine Website? (DE)](https://flize.io/wissen/wer-liest-deine-website.html): Google's share of verified bot traffic fell from 46.8% to 30.1% within a year (Cloudflare Radar, Q2 2025 vs. Q2 2026); new operators Anthropic, Apple and Baidu entered the top 9 with 13.2 points combined. What the shift toward answer-serving bots means for website visibility.
- [Who is reading your website? (EN)](https://flize.io/wissen/wer-liest-deine-website-en.html): English version.
- [Wie werde ich in Perplexity zitiert? (DE)](https://flize.io/wissen/wie-werde-ich-in-perplexity-zitiert.html): How Perplexity picks its citations (freshness, direct answers, dated facts), the role of PerplexityBot and Perplexity-User, and how Perplexity differs from ChatGPT. Sources: Perplexity docs.
- [How do I get cited by Perplexity? (EN)](https://flize.io/wissen/wie-werde-ich-in-perplexity-zitiert-en.html): English version.
- [Wie komme ich in Google AI Overviews? (DE)](https://flize.io/wissen/google-ai-overviews.html): How Google picks sources for the AI summary above search, what to optimise (direct answers, structured data, crawlability), AI Overviews vs. AI Mode. Sources: Google Search Central.
- [How do I get into Google AI Overviews? (EN)](https://flize.io/wissen/google-ai-overviews-en.html): English version.
- [Wie messe ich meine Sichtbarkeit in KI-Antworten? (DE)](https://flize.io/wissen/ki-sichtbarkeit-messen.html): Why rankings are not enough, the metrics that matter (mention rate, citation rate, context), how to collect data across engines over time, plus AI referral traffic.
- [How do I measure my visibility in AI answers? (EN)](https://flize.io/wissen/ki-sichtbarkeit-messen-en.html): English version.
- [Wie werde ich in ChatGPT zitiert? (DE)](https://flize.io/wissen/wie-werde-ich-in-chatgpt-zitiert.html): How ChatGPT picks sources, which content gets cited (direct answers, dated numbers, FAQ blocks), technical requirements (GPTBot access) and common mistakes.
- [How do I get cited in ChatGPT? (EN)](https://flize.io/wissen/wie-werde-ich-in-chatgpt-zitiert-en.html): English version.
- [GEO erklärt (DE)](https://flize.io/wissen/geo-erklaert.html): Why AI answers replace the ranking list: the two entry points into AI answers (trained model knowledge and live-retrieved sources) and how to become citable at both.
- [GEO explained (EN)](https://flize.io/wissen/geo-erklaert-en.html): English version.
- [Fragen-Scouting-Tools (DE)](https://flize.io/wissen/fragen-scouting-tools.html): Which tools find the questions users ask ChatGPT and Perplexity: AEO tools that query assistants with prompts, and tools that derive questions from a domain or brand name; why question scouting differs from keyword research.
- [Question scouting tools (EN)](https://flize.io/wissen/fragen-scouting-tools-en.html): English version.
- [Artikel, den KI zitiert (DE)](https://flize.io/wissen/artikel-den-ki-zitiert.html): How to write an article that AI systems cite: AI cites passages, not articles, so every paragraph must stand alone. Covers the answer block at the top, one question per H2 answered in the first sentence, numbers with source and date, FAQ blocks, and the structural mistakes that cost the citation (preamble, marketing speak, key content only in images or JavaScript).
- [Writing an article AI cites (EN)](https://flize.io/wissen/artikel-den-ki-zitiert-en.html): English version.
- [Fragen-Scouting für Versicherer (DE)](https://flize.io/wissen/fragen-scouting-versicherer.html): How question scouting differs from keyword research for insurers: keyword tools list terms and monthly search volume, question scouting captures whole customer questions with their intent, because a prompt is an instruction to generate text, not a relevance signal. Covers long-tail insurance questions, question-and-answer page structure, and how to run both research tracks side by side in an insurer's content process.
- [Question scouting for insurers (EN)](https://flize.io/wissen/fragen-scouting-versicherer-en.html): English version.
- [Reddit-Zitate in ChatGPT (DE)](https://flize.io/wissen/reddit-zitate-chatgpt.html): Reddit's share of ChatGPT Search citations held steady at 3.8% from July 18 to August 7, 2026, then fell below 1% on August 14 (0.52% average August 14-17), a relative drop of 86%; seven-day average 1.50%, 30-day average 3.15%. Source: PromptWatch, August 18, 2026, which names a change in source selection as the likely cause but cannot rule out a data-collection issue. Covers the methodology (measured through the AI interfaces, citation share counted only across responses with at least one source) and what platform-share volatility means for building visibility on your own domain.
- [Reddit citations in ChatGPT (EN)](https://flize.io/wissen/reddit-zitate-chatgpt-en.html): English version.

- [Warum werde ich nicht zitiert? (DE)](https://flize.io/wissen/warum-werde-ich-nicht-zitiert.html): A four-layer diagnosis for pages that never get cited in AI answers, to be worked through in order: reachable (robots.txt separates search crawlers from training crawlers, so blocking GPTBot is not the same as blocking OAI-SearchBot; CDN and firewall defaults such as Cloudflare's new Search/Agent/Training classes from September 15, 2026; Google-Extended does not govern AI Overviews), readable (no JavaScript execution, content in images, consent walls, wrong status codes), citable (answer format, passages that stand alone) and relevant (the long, case-bound questions people actually ask). Sources: OpenAI, Anthropic, Perplexity, Google Search Central, Cloudflare, all 2026.
- [Why am I not cited in AI answers? (EN)](https://flize.io/wissen/warum-werde-ich-nicht-zitiert-en.html): English version.
- [Welche KI-Crawler zulassen? (DE)](https://flize.io/wissen/welche-ki-crawler-zulassen.html): Which AI crawlers to allow and which to block, sorted by the three purposes providers now separate by bot: search (OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, Bingbot, Applebot, meta-webindexer) is the only purpose that produces citations with a link; training (GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, meta-externalagent, CCBot) is a separate decision; user-initiated fetches (ChatGPT-User, Claude-User, Perplexity-User, meta-externalfetcher) have a person behind them and may ignore robots.txt. Includes a full bot reference table with stated purpose and robots.txt behaviour, two ready-made robots.txt profiles, the rule that a bot reads only the group matching its name and ignores User-agent: * once named, and the three control layers (robots.txt asks, firewall decides, meta tags such as noarchive/nocache and nosnippet govern usage). Sources: OpenAI, Anthropic, Perplexity, Google Search Central, Apple, Meta, Bing Webmaster Blog, Cloudflare (training was nearly 80% of AI crawler traffic in July 2025; new Cloudflare defaults from September 15, 2026).
- [Which AI crawlers should I allow? (EN)](https://flize.io/wissen/welche-ki-crawler-zulassen-en.html): English version.
## Legal

- [Datenschutz (DE)](https://flize.io/datenschutz.html): Privacy policy (German).
- [Privacy (EN)](https://flize.io/datenschutz-en.html): Privacy policy (English).
- [Impressum (DE)](https://flize.io/impressum.html): Legal notice (German).
- [Legal Notice (EN)](https://flize.io/impressum-en.html): Legal notice (English).
- [Nutzungsbedingungen (DE)](https://flize.io/nutzungsbedingungen.html): Terms of use (German, legally binding version).
- [Terms of Use (EN)](https://flize.io/nutzungsbedingungen-en.html): Terms of use (English).

## Contact

- Email: kontakt@sitebrunch.com
- Company: sitebrunch GmbH, Eifflerstraße 43, 22769 Hamburg, Germany