The Glippy Extension for Chrome
The Glippy Extension grades any webpage for AI search and agent readiness in a single click. Run 240+ checks across 16 GEO categories, work in 31 languages, and layer on optional LLM-powered content analysis — all without leaving your browser.
What the Glippy Extension does
Install the Glippy Extension, open any page, and click the toolbar icon. In seconds you get a GEO readiness score from 0 to 100 with a letter grade, a per-category breakdown, and prioritized tips for your weakest areas. The score also lands right on the extension icon as a color-coded badge so you can audit while you browse.
- One-click scoring — instant GEO score and grade for the active tab
- Per-category detail — expand any of the 16 categories to see exactly which checks passed, warned, or failed
- Trend tracking — stores up to 50 scans per URL with a sparkline so you can see if your changes are working
- Competitor comparison — scan a rival URL side-by-side across all 16 categories
- Multi-page & bulk-tab scans — audit up to 10 URLs at once, or every open tab
- Exports (Pro) — HTML, PDF, Markdown, CSV, plain text, and direct upload to Google Drive
Privacy first: Glippy analyzes pages entirely in your browser. No page content is sent to Glippy servers — the only network calls are license verification and the optional LLM analysis you configure yourself.
All checks the Glippy Extension runs
Glippy evaluates 240+ signals grouped into 16 categories. Categories 1–10 score technical readiness; categories 11–16 score the content itself. Each category is weighted by its impact on how AI agents and generative search engines discover, understand, and cite your page. On top of these, version 2.3 adds an on-demand live AI crawler access check that probes your site as each AI crawler and reports what your server actually served.
Technical readiness (1–10)
🧬 1. Structured Data & Schema
JSON-LD presence, valid schema.org types and required properties, breadcrumbs, sameAs, and Speakable markup.
1.5× weight🏗️ 2. Semantic HTML
Heading hierarchy, single H1, landmark elements, lists and tables, and overall document outline quality.
1.2× weight♿ 3. Accessibility for Agents
ARIA labels and roles, image alt text, descriptive link text, form labels, and the html lang attribute.
1.0× weight🔗 4. Internal Linking
Link count and density, anchor-text quality, navigation structure, breadcrumbs, and orphaned links.
1.0× weight🔍 5. Meta & Discoverability
Title and description, Open Graph and Twitter cards, canonical, robots directives, robots.txt rules for LLM crawlers, llms.txt, and sitemap.
1.0× weight🤖 6. Machine Readability
SSR vs JS-only rendering, content-to-HTML ratio, readability, plus llms-full.txt, Content-Signal, Markdown sources, feed discovery, RSL licensing, and IndexNow.
1.5× weight🏛️ 7. Entity & Authority
Author and organization schema, bios and credentials, sameAs links, publication dates, and E-E-A-T signals.
1.0× weight💬 8. Citability & Answer-Readiness
FAQ blocks, direct-answer patterns, quote-worthy sentences, definitions, TL;DR sections, and step-by-step instructions.
1.3× weight⚡ 9. Performance & Crawlability
Page load estimates, DOM size, mobile viewport, Core Web Vitals indicators, and image optimization signals.
0.3× weight🔌 10. Agent Interactivity (WebMCP)
WebMCP declarative and imperative APIs (document.modelContext), machine-operable forms, SDK detection, plus agent-discovery surfaces like MCP/A2A cards, AP2 payments, NLWeb, schemamap, and Web Bot Auth.
Content quality (11–16)
🎯 11. Content Positioning
Brand differentiation language, proof points, social proof, and unique value propositions beyond generic claims.
1.2× weight📅 12. Content Freshness
Publication and modification dates, temporal references, and stale-date warning signals.
0.8× weight📊 13. Information Density
Substance-to-filler ratio, meaningful sections, claim-evidence pairing, and "foam" detection.
1.0× weight🔍 14. Factual Verifiability
Citations and source links, attribution for statistics, methodology disclosure, and links to primary sources.
0.8× weight📚 15. Content Comprehensiveness
Word count vs topic scope, heading coverage of subtopics, defined key terms, and edge cases.
0.8× weight🖼️ 16. Multimodal Content
Image alt text, figure/caption markup, video transcripts, accessible inline SVG, and VideoObject/ImageObject schema.
0.5× weightAI crawler access, measured live
The 16 categories above are graded from the page Glippy already has in your browser. New in version 2.3, the AI crawler access (live) panel does something different: on one click it sends 12 real requests to your site from your own network, all for the same URL, all identical except for the User-Agent, and reports what your server actually did with each one. It is the difference between reading your robots.txt and watching a crawler get turned away at the door.
🔎 8 AI crawlers probed
Search indexers (OAI-SearchBot, Claude-SearchBot, PerplexityBot), live fetchers (ChatGPT-User, Claude-User, Perplexity-User) and training crawlers (GPTBot, ClaudeBot), each with its published User-Agent.
🧪 4 control requests
A plain-browser baseline, Glippy's own honest bot identity, a spoofed Googlebot, and a closing baseline repeat. The controls are what turn a refusal into evidence.
⚖️ Difference, not status codes
A bare 403 proves nothing: it may only mean your IP has a poor reputation. Only a difference between these same-IP requests is treated as evidence about the site.
🛡️ Edge and WAF detection
Identifies the CDN or bot-protection vendor serving the refusal, and flags the case where your firewall and your robots.txt disagree.
⏱️ Fetcher timeout risk
Times the responses that were actually served, so you find out when a page is slow enough that a live fetcher gives up before it loads.
🤐 Says "unknown" when it is
If your plain browser request is refused too, or the site throttles the run, Glippy tells you the check could not conclude instead of reporting your site as hostile to AI.
Blocks are not all equal, and Glippy does not treat them that way. A blocked search indexer is the finding that matters, because those crawlers build the indexes AI answers cite. A blocked live fetcher is a lesser warning. A blocked training crawler is informational only: refusing model training is a legitimate publisher choice, and Glippy will never mark you down for it.
The check runs only when you click it, and Chrome asks for permission for that site the first time. It is free, works from any network, and takes about a minute. Glippy cannot send requests from OpenAI's or Anthropic's networks, and many firewalls decide on verified crawler IP rather than on User-Agent, so your own server access logs remain the ground truth. The panel says so on every result.
Can AI and LLMs access my page? → · Full live-check reference →
Languages the Glippy Extension supports
The Glippy Extension interface is fully localized into 31 languages and follows your browser language automatically. Glippy analyzes pages in any language — this list covers the languages the extension UI itself is translated into.
Using the LLM option
By default, Glippy scores the 16 categories with fast in-browser heuristics. The content categories (11–16) — positioning, freshness, information density, verifiability, comprehensiveness, and multimodal — are judgement calls that benefit from a real language model. That's what the LLM option is for.
The LLM option is a Pro, bring-your-own-key (BYOK) feature. When you turn it on and add your own API key, Glippy uses your chosen model to evaluate the content categories more deeply and folds them into your overall GEO score instead of keeping them as a separate Content score.
How to enable it
- Activate Glippy Pro in the extension (the LLM settings are Pro-only).
- Open Settings → LLM Analysis in the extension popup.
- Pick a provider and paste your own API key, then choose a model (a fast default is preselected).
- Click Test to validate the key, then Save. A status indicator confirms LLM analysis is active.
- Re-scan any page — the content categories are now LLM-assessed and counted toward your GEO score.
Supported LLM providers
You bring your own key for any one of six providers, so you stay in control of cost and model choice:
Anthropic Claude
Haiku 4.5, Sonnet 4.6, Opus 4.7 and more
OpenAI
GPT-4o mini, GPT-4o, GPT-5 mini, GPT-5
Google Gemini
Gemini 2.5 Flash & Pro, 3.x previews
Groq
Llama 3.3 70B, Llama 4 Maverick
DeepSeek
DeepSeek V3, DeepSeek R1
Mistral AI
Mistral Small, Mistral Large
Your API key is stored locally and used only to call the provider you select. With the LLM option off, Glippy still scores all 16 categories using its built-in heuristics — the key just unlocks deeper, model-graded content analysis.
Get the Glippy Extension
Install free from the Chrome Web Store and score your first page in seconds. No sign-up required.