25% of Gen Z and Millennial searches now begin on an AI chatbot (Bain, 2025). Only 12% of URLs cited by ChatGPT, Perplexity and Copilot rank in Google's top 10 for the same prompt (Ahrefs). If your GEO is broken, you are losing traffic you do not see in Analytics.
The free 60-second audit checks: whether top-tier AI models (GPT class, Claude, Gemini) actually know your brand, your llms.txt score, AI crawler access, Schema.org markup -- plus a screenshot of what AI sees on your homepage.
Run a free 60-second audit Audits from €25
Grounded in research from Ahrefs, Similarweb, Bain, Profound, Originality.ai, and the Aggarwal et al. GEO paper (KDD 2024). See the research.
Built by Torsten “Getty” Raudssus.
Go beyond the 60-second snapshot. Per-page citability, E-E-A-T scoring, platform readiness, and a prioritised cross-channel growth plan.
Twenty key pages, standard depth
Twenty pages crawled, ChatGPT/Claude/Perplexity/Gemini probed, scored report with concrete fixes.
Audit of up to 20 pages
25.00 EUR
See example audits
Twenty pages, top-tier reasoning
Same crawl footprint as Small Audit, but every strategic step runs on the top reasoning model.
Audit of up to 20 pages
With premium AI model per page
90.00 EUR
See example audits
Two hundred pages, standard depth
Deep crawl across two hundred pages, full per-page rewrites and strategic synthesis.
Audit of up to 200 pages
100.00 EUR
Learn moreBrand visibility across every AI language at once
Same twenty-page scope as the Small Audit, but your brand is probed in every supported language at once -- see where you surface in AI answers across markets, not just one.
Audit of up to 20 pages
95.00 EUR
No credit card. No signup. Paste a URL + email, get a quick GEO score across the major AI surfaces.
Please note: the free audit only scans your homepage and may produce ratings that differ from a full audit.
60-second AI visibility snapshot
Paste a URL. See where ChatGPT, Claude, Perplexity and Gemini source your brand.
Google ranked pages by backlinks and on-page SEO.
AI cites sources by brand mentions, entity graphs, and extractable passages.
Generative Engine Optimization is the practice of making your content cite-worthy to LLM-powered search engines. It requires three things traditional SEO doesn't:
Nine research-backed categories. Every score grounded in published studies, not vibes.
Score 0-100 for how likely ChatGPT, Claude, Perplexity and Gemini are to quote a passage. Based on the Aggarwal et al. GEO paper (arXiv 2311.09735 / KDD 2024): "Cite Sources", "Quotation Addition" and "Statistics Addition" rewrites lift visibility 30 - 40% across provider benchmarks, with rank-5 pages reaching 115% under "Cite Sources" specifically.
Scores your brand presence across the 5 platforms AI uses for entity recognition: Wikipedia, Reddit, YouTube, LinkedIn, News. Per-platform 0-100 with sentiment, subreddit quality, and news recency.
Parses robots.txt and classifies access for every major AI crawler: GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, Bingbot and six more.
Parses every JSON-LD block, validates against schema.org, audits sameAs linking across 10 canonical platforms (Wikipedia, Wikidata, LinkedIn, YouTube, X, Facebook, Crunchbase, GitHub, Instagram, TikTok).
Scores Experience, Expertise, Authoritativeness, Trustworthiness against Google's Search Quality Rater Guidelines (September 2025 revision -- expanded YMYL scope, added a dedicated AI Overviews evaluation chapter).
Scored separately per surface because the platforms genuinely diverge -- only ~11% of domains overlap between ChatGPT and Perplexity for the same query, ~21% between ChatGPT and AI Overviews (Profound).
Measures completeness, not just extractability. 0-100 across 6 dimensions: core concepts, use cases, implementation, comparison, pricing, latest trends.
Runs 5-7 test queries in your vertical against AI-citation-proxy endpoints: "For 'how to set up X', competitor 1 shows up in 67% of ChatGPT answers. You show up in 0%."
Not just an audit -- 5-8 prioritized actions from a curated playbook: citability rewrites, cold-email for earned mentions, Reddit community marketing, competitor-alternatives pages, programmatic SEO.
The free audit gives you a 60-second snapshot. The paid audit runs the full pipeline -- per-page citability, schema scan, E-E-A-T, per-platform readiness, and a prioritised growth plan. Page count (20 vs 200) and Pro premium AI model are sold as variants of the same paid audit, not separate products.
| Free Audit | Paid Audit | |
|---|---|---|
| Turnaround | Minutes | Minutes |
| Pages analysed | Homepage only | 20 (Small) or 200 (Big) |
| AI model | GPT-5.4-mini | GPT-5.5 (Standard) or GPT-5.5-pro (Pro) |
| llms.txt detection + score | ✅ Score 0-100 (Docs section + URL count) | ✅ Same score plus per-page coverage gaps and a concrete add/fix list |
| AI citability scoring | Score only | Per-page + domain aggregate |
| Schema audit | Score only | Per-page + aggregate |
| Cross-AI brand recall probe | ✅ 5 EU/US providers + 2 China providers on consent (DeepSeek, MiniMax) | ✅ Same 5 EU/US + 2 China providers on consent, with full citation + sentiment scoring |
| Findings + actionable plan | — | ✅ Full prioritised growth plan (5-8 actions) |
| Core Web Vitals (Lighthouse) | — | ✅ LCP / INP / CLS per page with priority fixes |
| E-E-A-T assessment | — | ✅ Experience / Expertise / Authoritativeness / Trust scored per page |
| Per-platform readiness | — | ✅ Google AI Overviews, ChatGPT Web Search, Perplexity, Gemini, Bing Copilot |
| Per-page copy rewrites | — | ✅ paste-ready answer-first openings, fact boxes, FAQ blocks |
| Report format | Online only |
PDF, DOCX, Markdown, JSON export (15-25 pages)
Small Audit Pro |
Every paid audit is an individually generated, fully automated report -- no personal consulting, no manual review, no human editorial work. Delivered as PDF, DOCX and a saveable online view.
Traditional SEO optimises for Google's ranking algorithm (backlinks, keywords, on-page content). GEO optimises for AI systems that cite sources in generated answers. The signals differ: in the Ahrefs 75,000-brand study (2025), YouTube mentions correlate with AI citation at 0.737 while Domain Rating correlates with ChatGPT citation at only 0.266 -- the brand-mentions signal beats the link signal for that pair. GEO does not replace SEO, it runs in parallel and requires different signals.
Three concrete differences:
Almost every product fans out across many models -- the audit calls a wide spread of providers (OpenAI, Anthropic, Google, Perplexity, DeepSeek, MiniMax) to cross-check what each one actually says about your brand. The qualitative heavy lifting (E-E-A-T judgement, growth-plan synthesis, finding-by-finding reasoning) lands on one of the current top-tier model series -- the GPT-4/5 line, the Claude 4 line or the Gemini 2.x line, depending on the step. We deliberately name model series and not specific model versions: provider catalogues shift every few weeks, and the audit's quality target is the series, not a frozen version string.
Yes -- and we don't pretend otherwise. LLMs are not deterministic on every sub-question, and different model series weight evidence slightly differently. We mitigate this with strict tool schemas, score formulas computed in Perl (not by the LLM), and aggregation across multiple providers for the AI-visibility section -- but a re-run can still move category scores by a few points. We optimise for the directional signal ("this category is weak, here's exactly what to fix"), not for a stable three-decimal number.
No. Agent loops give one model a goal and a tool belt and let it decide its own steps -- they drift, hit context limits, and lose the original objective on long tasks. This audit is the opposite: a fixed, ordered pipeline of small, scoped steps. Each step gets a tight system prompt, a strict tool schema, and only the inputs it needs. The LLM judges one well-framed question at a time; the orchestration, the scoring, and the report assembly are deterministic Perl. That's why the output is reproducible in shape, even when individual judgements wobble.
No. Your content is fetched once via our self-hosted Crawl4AI instance, analysed, and used only to generate your report. We do not feed your content to any third-party training pipeline. You can set noai or noimageai meta-robots tokens without affecting the audit.
Yes, as long as our crawler can reach the URL. And it works better than people expect: GEO is mostly about passive on-page signals (schema, citability of copy, crawler permissions) plus what AI models already know about your brand name. The brand-knowledge signal keys off the company name, not the domain -- so if the staging copy and schema match what production will ship, the audit values shouldn't shift much after the deploy. The report flags itself as a staging run so you remember it's a pre-launch snapshot.
We won't tell you to do it on a fixed schedule -- we earn from these audits and that would be the obvious answer. The honest one: re-running makes sense whenever the AI market takes a noticeable step, e.g. a new GPT or Claude generation lands, Google flips an AI Overviews behaviour, a major schema standard changes. At those moments the model that scores you and the models that decide whether to cite you have both moved -- so the picture can move too. Between such shifts, on a stable site, scores tend to drift slowly. How often that's worth re-checking depends on how strategic GEO traffic is for your business.
In about a minute the free audit delivers:
The deep qualitative steps -- E-E-A-T judgement, per-page copy rewrites, a growth plan, multi-provider AI-visibility cross-checking -- each take several minutes of LLM compute and cost us real money per run. Giving all of that away without a credit card would not be sustainable. The free tier is designed so that the checks we do give are cheap to run (mostly deterministic HTTP probes and one fast brand-recall pass) but still directionally useful: you see immediately whether AI knows your brand and what the biggest blockers are.