Free GEO Audit Tool
Quick answer: a GEO audit (Generative Engine Optimization audit) shows whether AI assistants such as ChatGPT, Claude and Perplexity can reach, read and quote your website. Type your address below and in about 15 seconds you get a score from 0 to 100, a list of what to fix first and a PDF report. The tool runs up to 81 checks on your homepage and up to 8 inner pages. It is free and needs no sign-up or email.
GEO Audit at a Glance
| Question | Answer |
|---|---|
| What it is | A free check of how ready a website is for AI search and AI assistants |
| Checks | Up to 81 checks in 8 categories. 72 of them count towards the score, the rest are for information |
| Pages tested | The homepage and up to 8 inner pages sampled from your sitemap |
| AI crawlers | 25 crawlers checked in robots.txt, 10 of them also tested with a live request |
| Time | About 10 to 20 seconds |
| Cost | Free. No sign-up, no email, 5 audits per hour |
| What you get | A score from 0 to 100, a score per category, the five fixes that matter most and a PDF report |
| Scoring | Published in full on the GEO audit methodology page |
How to Run a GEO Audit
- Enter your website address. Type the domain into the box above and press Check my website. The audit always starts from the homepage.
- Wait about 15 seconds. Our server fetches your homepage, robots.txt, sitemap and llms.txt, requests the homepage again as 10 AI crawlers, and reads up to 8 inner pages.
- Read the score. You get one score from 0 to 100 and a separate score for every category, so you can see where the site is weakest.
- Fix the top five first. The Fix these first list names the five changes with the biggest effect. Many of them link to a free tool that goes deeper.
- Download the PDF and test real answers. The report ends with six prompts to paste into ChatGPT, Claude and Perplexity, so you can see what they say about your site.
What the GEO Audit Checks
The audit runs up to 81 checks in 8 categories. Some checks only apply to certain sites, for example article checks need pages with Article markup, so the exact number depends on your site.
| Category | Checks | What it looks at |
|---|---|---|
| AI readiness | 20 | Whether AI crawlers are allowed in and get a readable page: robots.txt rules by crawler type, a live firewall test, content without JavaScript, freshness, trust pages, author and source signals on articles. |
| Content & citability | 12 | Whether the text is easy to lift into an answer: a summary near the top, sections that open with a short answer, definitions, lists and tables, sourced statistics, sentence and paragraph length, hidden instructions aimed at AI. |
| Indexing & crawling | 11 | robots.txt status, HTTPS and certificate expiry, one version of the site, canonical URL, crawler-specific noindex, snippet limits, redirect chain, hreflang, XML sitemap and real 404s. |
| On-page content | 9 | Title tag, meta description, a single H1, heading order, image alt text, language, mobile viewport, Open Graph tags and favicon. |
| Structured data | 8 | JSON-LD syntax, Organization and WebSite markup, logo, linked profiles (sameAs), one consistent brand name, and the key fields of ten common schema types. |
| Links & pages | 9 | Internal links, link text and broken links, plus up to 8 pages sampled from your sitemap: status, titles, descriptions, H1, duplicates, thin content and canonical tags. |
| Rendering & speed | 6 | Server response time, HTML size, compression, text-to-code ratio, number of elements and render-blocking scripts. |
| AI standards (optional) | 6 | llms.txt and llms-full.txt, Content-Signal lines, noai tags, a Markdown version and an RSS feed. Reported for information only and not scored. |
Every check, its weight and the scoring rules are listed on the GEO audit methodology page.
AI Crawlers We Test
The audit reads your robots.txt rules for 25 crawlers and sorts them into four types, because blocking each type has a different effect. 10 of them are also tested with a live request, to see whether a firewall or CDN turns them away.
| Crawler | Company | Type | Used for | Live test |
|---|---|---|---|---|
| OAI-SearchBot | OpenAI | AI search crawler | Indexes pages for AI answers | Yes |
| Claude-SearchBot | Anthropic | AI search crawler | Indexes pages for AI answers | Yes |
| PerplexityBot | Perplexity | AI search crawler | Indexes pages for AI answers | Yes |
| Googlebot | Search engine | Builds the index behind Google AI Overviews, Gemini and Copilot | Yes | |
| Bingbot | Microsoft | Search engine | Builds the index behind Google AI Overviews, Gemini and Copilot | Yes |
| Applebot | Apple | AI search crawler | Indexes pages for AI answers | Yes |
| DuckAssistBot | DuckDuckGo | AI search crawler | Indexes pages for AI answers | Yes |
| MistralAI-Index | Mistral | AI search crawler | Indexes pages for AI answers | robots.txt only |
| Meta-WebIndexer | Meta | AI search crawler | Indexes pages for AI answers | robots.txt only |
| Amzn-SearchBot | Amazon | AI search crawler | Indexes pages for AI answers | robots.txt only |
| ChatGPT-User | OpenAI | User-triggered fetcher | Opens a page when a person asks the assistant about it | Yes |
| Claude-User | Anthropic | User-triggered fetcher | Opens a page when a person asks the assistant about it | Yes |
| Perplexity-User | Perplexity | User-triggered fetcher | Opens a page when a person asks the assistant about it | Yes |
| MistralAI-User | Mistral | User-triggered fetcher | Opens a page when a person asks the assistant about it | robots.txt only |
| Meta-ExternalFetcher | Meta | User-triggered fetcher | Opens a page when a person asks the assistant about it | robots.txt only |
| Amzn-User | Amazon | User-triggered fetcher | Opens a page when a person asks the assistant about it | robots.txt only |
| GPTBot | OpenAI | Training crawler | Collects text for model training | robots.txt only |
| ClaudeBot | Anthropic | Training crawler | Collects text for model training | robots.txt only |
| Google-Extended | Training crawler | Collects text for model training | robots.txt only | |
| Applebot-Extended | Apple | Training crawler | Collects text for model training | robots.txt only |
| CCBot | Common Crawl | Training crawler | Collects text for model training | robots.txt only |
| Bytespider | ByteDance | Training crawler | Collects text for model training | robots.txt only |
| meta-externalagent | Meta | Training crawler | Collects text for model training | robots.txt only |
| Amazonbot | Amazon | Training crawler | Collects text for model training | robots.txt only |
| cohere-ai | Cohere | Training crawler | Collects text for model training | robots.txt only |
Blocking a training crawler such as GPTBot or ClaudeBot is a policy choice. It does not remove your site from AI answers, so the audit reports it without taking points off. For a crawler-only report, use the AI Crawler Checker.
How the Score Is Calculated
Every scored check has a weight from 1 to 3. A passed check earns its full weight, a warning earns a quarter and a failed check earns nothing. Checks measured across several pages or images, such as alt text or author bylines, earn credit in proportion to how many pass. The score is the points you earned divided by the points available, shown as a number from 0 to 100. Each failed check with the top weight of 3 takes off 5 more points, up to 25. The score measures AI readiness: whether AI search engines can reach, read and quote the site. It does not measure whether assistants already mention you, so the report also gives you prompts to check that yourself.
| Score | Meaning |
|---|---|
| 80 to 100 | Good. Only small fixes are left. |
| 50 to 79 | Needs work. Fix the red items first, then the amber ones. |
| 0 to 49 | Poor. Crawlers or visitors are likely running into real problems. |
One rule overrides the rest. If something stops AI engines from using the site at all, the score is held at 35 or lower until it is fixed. That covers a homepage that does not answer 200, a noindex directive, robots.txt blocking Googlebot, Bingbot or most AI search crawlers, a firewall that turns most AI crawlers away, and a page with almost no text before JavaScript runs. The full rules are on the methodology page.
GEO Audit vs SEO Audit
A GEO audit asks whether AI assistants can reach, read and quote a page. An SEO audit asks whether a search engine can crawl, index and rank it. The two overlap on the basics, and each adds checks the other does not need.
| GEO audit | SEO audit | |
|---|---|---|
| Main question | Can AI assistants reach, read and quote this site? | Can search engines crawl, index and rank this site? |
| Crawlers | 25 crawlers, including AI search crawlers and user-triggered fetchers | Googlebot and Bingbot |
| Content | Short answers, definitions, sourced statistics, text without JavaScript | Titles, descriptions, headings, thin and duplicate pages |
| Trust | Author, dates, sources, one consistent brand entity | Security headers, HTTPS, mixed content |
| Speed | Response time and HTML weight, because AI fetchers give up quickly | Caching, image weight, lazy loading, render-blocking files |
| Tool | This page | Free SEO Audit |
Common GEO Audit Failures and How to Fix Them
AI crawlers blocked in robots.txt
A Disallow rule for OAI-SearchBot, Claude-SearchBot or PerplexityBot keeps your pages out of the index those assistants answer from. Remove the rule, or add an Allow line for the crawler. Test your file with the Robots.txt Tester.
A firewall or CDN turning AI crawlers away
robots.txt can allow a crawler while the firewall answers it with 403. This often comes from a “block AI bots” switch in a CDN such as Cloudflare. Allow the answer crawlers in your bot rules.
Content that only appears after JavaScript runs
Most AI crawlers do not run JavaScript, so a page built in the browser looks empty to them. Render the main content on the server or pre-render it. You can see the raw HTML with View Page Source.
A noindex or snippet limit left on the page
A noindex directive, or one aimed at a single crawler such as bingbot, removes the page from every product built on that index. max-snippet:0 and nosnippet stop engines from quoting the page. Remove them unless you mean them.
Sections that do not open with an answer
AI assistants quote short, self-contained passages. Answer each heading in the first paragraph below it, in 15 to 80 words, and name the subject instead of starting with “it” or “this”. Add the detail after that.
Statistics with no source
A number without a source is hard for an assistant to trust. Name or link the source next to every statistic.
How This Tool Compares
Several free GEO audit tools exist, and they measure different things. This table compares three of them using what each tool states on its own page (checked in November 2026).
| MaslonLabs GEO Audit | Juma GEO Audit | Feedough GEO Audit | |
|---|---|---|---|
| Price and sign-up | Free, no sign-up | Free, no sign-up | Free, no login |
| Checks | Up to 81 in 8 categories | 10 weighted dimensions | About 85 in 8 categories |
| Pages per audit | Homepage and up to 8 inner pages | Not stated | One page |
| Crawlers in robots.txt | 25 | 7 | 27 |
| Live firewall test | 10 AI user agents | Not listed in its methodology | 6 AI user agents |
| Scoring method | Published, with every check and weight | Published, with weights per dimension | Explained on the tool page |
| Report | On-page report and a PDF download | Shareable public link | On-page report, Chrome extension |
| Also offers | Broken links, duplicate titles and thin pages across the sample | Brand presence in search results | A test of real searches in Brave and Google |
Pick by what you need. This tool is the one to use when you want a multi-page sample and a PDF to hand to a client or developer.
What You Get
- A short Fix these first list with the five changes that matter most, each with a link to the free tool that goes deeper
- A score from 0 to 100 and a score per category
- A table of the pages we sampled
- A PDF report you can download and share
Test It Yourself
The report ends with six ready-made prompts to paste into ChatGPT, Claude and Perplexity, so you can see what they say about your site. This is a manual check, because the audit cannot see what AI assistants answer.
What This Audit Does Not Measure
This is a readiness check run from one server location. It does not test whether ChatGPT, Claude or Perplexity actually mention your brand in their answers, and it does not measure rankings, backlinks or Core Web Vitals. The content checks follow best practice for quotable writing; no AI engine publishes them as rules. A high score means AI crawlers can reach and understand your site. Whether AI tools cite you also depends on your content and how well known your site is. Looking for classic Google issues? Try the free SEO audit.
How Often Should You Run a GEO Audit?
Run a GEO audit after any change to your theme or templates, robots.txt, CDN or firewall settings, and once a month for the sites you care about. Crawler rules and bot-blocking defaults change without notice, so a site that passed in spring can be blocked by autumn.
Frequently Asked Questions
What is a GEO audit?
A GEO audit checks how well a website can be discovered, read and trusted by AI search tools such as ChatGPT, Claude, Perplexity and Google AI Overviews. It looks at crawler access, page structure, structured data, content that is easy to quote and trust signals.
What is a good GEO score?
A score of 80 or more is good: AI crawlers can reach and read the site and only small fixes are left. Between 50 and 79 there are real gaps to close. Below 50 something is stopping crawlers or readers. A score held at 35 means one problem blocks AI engines completely and has to be fixed first. A high score means the site is ready for AI search, not that assistants already recommend it.
Is SEO replaced by GEO?
No. GEO builds on SEO. AI tools can only use pages they can crawl and understand, so the technical and content basics still matter. A GEO audit adds the checks that are specific to AI crawlers.
Should I block GPTBot?
That is your choice, and it does not remove you from ChatGPT answers. GPTBot collects text for training OpenAI models. ChatGPT search uses a different crawler, OAI-SearchBot, and ChatGPT-User opens pages when a person asks for them. The audit reports training crawlers separately and never takes points off for blocking them.
Do I need an llms.txt file?
No. llms.txt is optional, and no large AI engine has confirmed that it reads the file. The audit checks its format if you have one, but it does not change your score.
Is the audit really free?
Yes. There is no sign-up and no email. You can download the PDF report straight away.
What data do you save?
We save the audit type, the audited domain, the date, the score and your country, so we can see how the tool is used. The country is worked out from your IP address, which is not stored. We do not ask for your email. For the hourly limit we keep a temporary hash of your IP address for one hour, not in the log.
Will a high score get my site cited by AI assistants?
No tool can promise that. The audit shows whether AI crawlers can reach and read your pages. Being cited also depends on the quality of your content and your site’s reputation.
Why is there a limit of 5 audits per hour?
Each audit fetches many pages from your site. The limit keeps the tool free and available for everyone.
Need help fixing what the audit found? See our SEO & GEO service or contact MaslonLabs.
