SEO and GEO, one line at a time.
The terms behind getting found on Google and cited by AI, in plain words. Each one links to the check that protects it.
Terms
- GEO (generative engine optimization)#
- Being trusted and cited as a source when people ask AI assistants such as ChatGPT, Claude, Perplexity or Gemini. It depends on the whole site: crawler access, a clear brand entity and claims that link their sources.GEO guide
- AEO (answer engine optimization)#
- Being picked as the direct answer: the passage a featured snippet, a Google AI Overview or a voice assistant quotes. It depends on single passages: a question heading, the answer right below it, a date and a named publisher. GEO is the wider job of being a trusted source.GEO guide
- AI SEO, LLM SEO#
- Umbrella terms for making a site show up in AI-generated answers. In practice they cover both GEO and AEO, on top of solid technical SEO.GEO guide
- Google AI Overviews#
- AI-written summaries Google shows above some search results, with links to the pages they draw on. They come from Google's search index, so a page Googlebot can't crawl or index can't appear in them. Paradigma tracks how often they appear for your queries and whether you're in them.Search engines blockedPages turned noindex
- AI crawlers#
- Bots AI companies use to read the web. Search crawlers fetch pages so assistants can show and cite them; training crawlers collect data for models. robots.txt can allow one kind and block the other.AI search crawlers blockedAI training crawlers blocked
- GPTBot#
- OpenAI's crawler for model training. Blocking it opts out of training; it doesn't stop ChatGPT search from citing you, which uses OAI-SearchBot.AI training crawlers blocked
- OAI-SearchBot#
- OpenAI's search crawler: it fetches pages so ChatGPT search can show and cite them. ChatGPT-User is the agent that opens pages while answering a user.AI search crawlers blocked
- ClaudeBot#
- Anthropic's crawler for model training. Claude-SearchBot builds Claude's search results and citations, and Claude-User opens pages while Claude answers a user.AI training crawlers blocked
- PerplexityBot#
- Perplexity's crawler for its search index and citations. Perplexity-User opens pages while Perplexity answers a user.AI search crawlers blocked
- Google-Extended#
- A robots.txt token, not a separate crawler: it controls whether Google may use your content for Gemini training and grounding. It doesn't affect Google Search or AI Overviews.AI training crawlers blocked
- robots.txt#
- A file at the root of your site that tells crawlers which paths they may fetch. In Next.js you can write it as app/robots.ts.Search engines blocked
- Mention and citation#
- A mention is when an AI answer names your brand. A citation is when it links to one of your pages as a source. Both are measured as a rate over many sampled answers.GEO guide
- Core Web Vitals#
- Google's three real-user speed metrics: LCP (how fast the main content appears, good at 2.5 s or less), INP (how fast the page reacts, good at 200 ms or less) and CLS (how much the layout jumps, good at 0.1 or less).LCP regressionsINP regressionsCLS regressions
- CrUX (Chrome UX Report)#
- Google's public dataset of real Chrome users' experience on popular sites, reported over rolling 28-day windows. It is where field Core Web Vitals come from.Core Web Vitals
- Canonical URL#
- The URL a page declares as its main address, so search engines index one version instead of its variants. In Next.js it is alternates.canonical in the metadata.Canonical URL changed
- noindex#
- A directive, in a meta tag or the X-Robots-Tag header, that asks search engines to keep a page out of their results. In Next.js it is robots: { index: false }.Pages turned noindex
- Sitemap#
- An XML list of the URLs you want crawled, usually at /sitemap.xml. In Next.js you can generate it with app/sitemap.ts.URLs droppedSitemap removed
- hreflang#
- Links that tie the language versions of a page together, so each audience gets its own version in search.hreflang alternates
- Schema markup (structured data, JSON-LD)#
- Machine-readable facts about a page, such as a product, an article or an FAQ, written in the schema.org vocabulary. JSON-LD is the script format Google recommends.Structured data removedInvalid JSON-LD
- llms.txt#
- A short Markdown file at /llms.txt that summarizes a site and links to the pages that matter, for AI agents and developers working with it. Google Search doesn't use it.Our llms.txt
- MCP (Model Context Protocol)#
- An open protocol that lets AI assistants such as Claude or ChatGPT use tools and data from other services. Paradigma's MCP server lets your assistant read your SEO and GEO data for the projects you approve.Paradigma for AI tools
- MCP server card#
- A small JSON file at a well-known URL that describes a site's MCP server, so AI agents can discover what it offers.Agent readiness
- Agent readiness#
- How well AI agents can read, navigate and act on your site: content in the HTML, clear titles and language, crawler access, llms.txt and machine-readable entry points.Agent readiness
- Server-side rendering#
- Producing a page's HTML on the server, so crawlers and agents that don't run JavaScript still see the content. Next.js Server Components do this by default.Content lost on the Preview
- Google Search Console#
- Google's free tool that reports how your site performs in Google Search: queries, clicks, impressions, click-through rate and average position.Search click drops
- Click-through rate (CTR)#
- Clicks divided by impressions: the share of people who saw your result and clicked it.Low CTR
- Preview deployment#
- A temporary deployment of a branch or pull request, such as a Vercel Preview. Paradigma checks the pages it serves before you merge.Preview checks
Where to go next
- GEO guide: how AI answers pick sources, and how Paradigma tracks yours.
- Next.js SEO guide: what breaks rankings in App Router apps.
- All rules: every check, why it matters and how to fix it.