Free GEO Audit: Check Access First, Then Accuracy

A GEO audit answers two questions. Can AI engines get to your content? And when they answer questions about you, are the facts right?

The first question has strong evidence behind it. The second is the one customers actually notice. Everything else sold as GEO sits on weaker ground, as the critical look at lab GEO studies versus live engines explains.

It helps to separate four stages:

• Parametric: what a model knows with no web access. Nobody controls this in the short term.
• Access: whether a crawler may fetch and store your page. You and your CDN control this.
• Retrieval: whether your page enters the model's context for a query. The engine controls this, mostly.
• Generation: which retrieved pages get used and cited. The model controls this.

A free audit lives mostly in the access stage, because that is where you can find something broken and fix it.

Work through these in order. Stop and fix anything broken before moving on.

1. Read your robots.txt per bot. Run curl -s https://yoursite.com/robots.txt. Check each retrieval bot separately: OAI-SearchBot (ChatGPT search), PerplexityBot, Claude-SearchBot, Meta-WebIndexer and Googlebot. OpenAI's crawler page says each setting is independent, so you can block training and still allow search:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

2. Test the firewall, not just the file. A CDN or firewall can block a crawler that robots.txt allows. Request a key page with each crawler's user-agent string, copied from the operator's documentation: curl -s -o /dev/null -w "%{http_code}" -A "<user-agent string>" https://yoursite.com/pricing. A 403, 429 or challenge page is a failure. Your test comes from your own IP, so also check your CDN's bot settings. Cloudflare changed its AI crawler defaults for some accounts on 15 September 2026. The AI crawler access audit playbook covers this step in depth.

3. Diff raw HTML against the rendered page. Run curl -s https://yoursite.com/pricing | grep -i "your price". If a fact shows in your browser but not in the raw HTML, it depends on JavaScript. Vercel found ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. Googlebot renders, so this matters for non-Google engines.

4. Check index and snippet eligibility. Google says a page must be indexed and eligible for a snippet to appear as a supporting link in AI Overviews or AI Mode. Search your templates for noindex, nosnippet, max-snippet:0 and data-nosnippet. In Search Console, confirm the page is indexed and look for the generative AI inclusion control if your property has it. Inclusion is the default, but child properties can inherit an opt-out.

5. Check factual accuracy. Ask each engine three or four questions a buyer would ask. Note every wrong price, product name or founding fact. Then check your own site for the same fact stated differently on different pages.

Many GEO audits score you on adding quotations, statistics, authoritative-sounding phrasing or keyword variants. The evidence points the other way.

C-SEO Bench, a NeurIPS 2025 benchmark, found that most current conversational SEO methods are "largely ineffective" and "frequently have a negative impact on document ranking." Those tests ran on documents already inside the model's context, so they say nothing positive about getting found in the first place. Ahrefs' matched-control test found AI Overview citations fell 4.6% after pages added schema.

So a good audit flags heavy rewriting as a risk. It does not tell you to remove existing markup either: the honest line is that it is not a lever for AI answers. The comparison of rewrites and access fixes lays out both sides.

Injected content, such as hidden instructions aimed at AI crawlers, is a separate category. It is manipulation, it can backfire, and it belongs on the risk list only.

Sort your findings into three buckets.

• Broken: a retrieval bot disallowed, a 403 from the firewall, key facts only in JavaScript, or a stray noindex. Fix these first. They are binary and fully in your control.
• Wrong: an engine states a fact about you that is false. Make sure the correct fact appears plainly in fetchable HTML, stated once and consistently. A specific sentence, like a price or a one-line definition, gives an engine something it can quote.
• Out of your hands: whether an engine retrieves you for a given query, and which sources it cites. When engines cite other sites instead of yours, being named on those sites is associated with appearing in answers.

If you're unsure what to tackle first, the case for fixing crawler access before tracking is short and clear.

A hand audit is a snapshot of one day and a handful of questions. It has real blind spots.

• Engines rewrite your question into their own search queries, so your test prompts are a proxy.
• Retrieval is discretionary. Some answers come from the model's frozen knowledge, which you can't change this quarter.
• Google's Search Console AI report shows impressions only, with no position data. Any "AI rank" number has nothing to be checked against.
• Results differ by engine. A single blended score hides that, so record each platform separately.

The audit tells you what blocks you. It can't promise that fixing those blocks earns a citation.

If you want a structured list to work from, 99Visibility's free AI Content Checklist is the matching tool for this audit. Use it alongside the AI SEO checklist for the access and measurement steps.

For step five, the free checker on this site asks up to three buyer questions and shows whether AI engines name you, as a snapshot rather than ongoing tracking. For engine-specific detail, see the ChatGPT visibility checklist, or compare options in free GEO tools.

Does allowing GPTBot help me show up in ChatGPT? No. OpenAI's crawler documentation says GPTBot governs training use only. OAI-SearchBot controls whether your site can appear in ChatGPT search answers. The two settings are independent, so you can block GPTBot and allow OAI-SearchBot. If you want ChatGPT search eligibility, check OAI-SearchBot in robots.txt and make sure your firewall lets it through.

Should I add schema markup as part of a GEO audit? Not as a fix for AI answers. Google says structured data isn't required for generative AI search, and Ahrefs' matched test found AI Overview citations fell 4.6% after pages added schema. That test only covered pages already cited heavily. Existing markup does no harm, so there's no reason to strip it out. Just don't count it as an AI visibility lever.

How often should I rerun a free GEO audit? Rerun the access checks after any change to your CDN, firewall, robots.txt, templates or front-end framework. Those are the changes that silently break crawler access. Some access changes happen without you, such as a CDN changing its defaults. Rerun the accuracy questions when you change pricing, products or positioning, since that's when engines are most likely to repeat stale facts.

Why does my firewall matter if robots.txt allows AI crawlers? robots.txt is a request, and the firewall is the actual gate. OpenAI's help documentation says eligibility for ChatGPT search needs both: OAI-SearchBot allowed in robots.txt and your host or CDN allowing traffic from OpenAI's published IP ranges. A firewall returning 403 to a crawler removes you quietly, and nothing in your analytics will flag it.

Can a free audit tell me my AI ranking? No, and be wary of any tool that claims to. Google's AI reporting in Search Console exposes impressions only, with no position data. Other engines don't publish ranking data either. A free audit can tell you whether crawlers can reach and read your pages, and whether answers state your facts correctly. Those are measurable. A rank inside an AI answer is not.