GEO Insights & llms.txt
How ready your site is to be found, understood and cited by AI search and assistants. Computed from your own content and robots.txt, with no API key or AI calls.
The report is rebuilt every time the Content Audit runs. Until then only the crawler table (checked live) is shown.
- Generative-search readiness card and its parts
- AI Crawlers & Agents table
- Entities table
Generative-search readiness
A score out of 100 (70+ good, 40–69 needs work), from five parts:
| Part | Weight | What it measures |
|---|---|---|
| Site signals | 25 | About, contact and privacy pages, tagline, AI crawler access and other site-wide checks. |
| Content GEO score | 30 | Average GEO score of audited entries. |
| AI search access | 20 | Share of AI search and assistant crawlers robots.txt lets in. |
| Machine-readable content | 15 | Share of entries whose structured data says what they are (article, product, FAQ…). |
| Name consistency | 10 | Fewer spellings of the same name score higher. |
History keeps one snapshot per audit day (readiness, SEO, GEO, AEO, entries); the last 10 are shown here, with trends on the Dashboard.
AI crawler access
What your robots.txt (the real file, or the one WordPress generates with every plugin's changes) lets each AI crawler read, tested on your home page and newest post.
| Purpose | Crawlers |
|---|---|
| AI search | OAI-SearchBot (ChatGPT search), Claude-SearchBot, PerplexityBot, Googlebot (Search and AI Overviews), Bingbot (Bing and Copilot), Applebot (Siri and Spotlight), DuckAssistBot |
| On request | ChatGPT-User, Claude-User, Perplexity-User |
| Training | GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Amazonbot, CCBot, Bytespider |
Blocking training crawlers doesn't affect AI search; blocking search or on-request crawlers keeps you out of those answers. A blocked search crawler is flagged in the free Health Checks. To block only training crawlers, use the free AI training opt-out.
Entities & consistency
- Entities: the brands, products, places and people your content names (in 2+ entries), which page is about each, and what they appear with. Names mentioned in 3+ entries without a page of their own are listed: a page that defines each gives AI answers something to cite.
- Brand & Entity Consistency: names written more than one way (e.g. "WooCommerce" and "Woocommerce"). Your brand is flagged first. Pick one spelling and use it everywhere.
Topic authority & opportunities
- Topic Authority: for each topic cluster, a score from the number of entries, their depth (average words), how well they interlink and their GEO scores, with the pillar page.
- GEO Opportunities: entries with a GEO score under 80, ranked by how many pages link to them, with the GEO checks they miss, plus the most common GEO gaps site-wide.
- Machine-Readable Content: how many entries carry a specific schema type (not just WebPage), the most used types, and examples of generic-only entries.
llms.txt
A short, plain-text map of your important pages for AI assistants, served at your-site.com/llms.txt. Turn on llms.txt under Pro Features (off by default), then Crawl & Indexing → llms.txt.
One section per content type.
Pages are listed in menu order, other types newest first.
One or two sentences about the site. Empty uses your tagline.
# Your Site > Handmade coffee gear and brewing guides from Lisbon. ## Pages - [About us](https://your-site.com/about/): Who we are and how we roast. - [Shipping & Returns](https://your-site.com/shipping-returns/): Delivery times and our 30-day returns. ## Posts - [How to brew with a French press](https://your-site.com/french-press/): A 4-minute method. ## Optional - [XML sitemap](https://your-site.com/wp-sitemap.xml): Every public page of the site.
- Only published, indexable entries without a password, with their meta description (or excerpt).
- Refreshed when content changes, and served with
X-Robots-Tag: noindex. - A real
llms.txtfile on your server always takes precedence. Crawlers look for it at the domain root only. - Set Cart, Checkout and My Account to noindex (or leave Pages unticked) if you don't want them listed.