How to use
CiteLens

Type or paste the full page URL into the top field and click Analyze. Best for quick audits of public pages.

Use this when the URL method fails, or for pages behind logins and staging environments. Works on every page.

The analyzer retrieves pages through a secure proxy. Some sites like e-commerce, news, or automotive platforms actively block automated requests as a security measure on their end. When that happens, the HTML paste method is the solution: open the page in your own browser, copy the source, and paste it in. Your browser already has the full page we just analyze what you give us.

When to use each method:

• URL analysis works for most marketing pages, blogs, documentation, and product pages on standard sites.
• HTML paste use for any site that blocks the URL method, pages behind login, staging environments, or JavaScript-heavy sites where the URL fetch returns incomplete content.

A GEO score only tells part of the story. A score of 55% looks very different if your top competitor scores 45% vs 70%. The comparison puts your score in context — and changes what "good enough" actually means for your specific competitive landscape.

Pages are ranked by their overall GEO score — a weighted average across all four categories. A competitor with a high score in one category (e.g. AI Trust: 90%) may still rank below you overall if they score poorly in Answer-Readiness or E-E-A-T. The insight panel at the bottom of the comparison explains your position and flags any category where competitors are ahead.

Run the analyzer on every page before it goes live. Make it part of the final review checklist alongside SEO and proofreading. A 5-minute analysis before publishing is far easier than retrofitting a live page.

The action plan is the first thing you see in the What to Fix tab — ordered by priority. Don't try to fix everything at once. Resolve all High Impact items first — these typically move the score by 10–20 points on their own.

Download the report after each analysis and save it with the date in the filename. This gives you a record of GEO progress per page over time — useful for reporting to stakeholders.

Open a competitor's page, copy the HTML source, and paste it into the analyzer. You'll see exactly where they score well and where their gaps are — useful for identifying content opportunities.

The three things that move blog scores the most: a named author with credentials, a visible publication date, and at least 3 outbound links to credible sources. Make these standard in your blog template.

Focus on proof ratio — the percentage of claims backed by evidence. Generic claims like "enterprise-grade" or "industry-leading" hurt your score. Replace them with specific, verifiable statements backed by data or customer examples.

Pages in these languages are fully scored across all checks. The analyzer recognizes verbs, question words, research references, CTA phrases, and content patterns in each language — so the analysis is as accurate as it would be for an English page.

Right-to-left language pages are detected automatically — either from the page's language tag or by scanning the text itself. All structural checks are fully scored: schema, headings, word count, links, HTTPS, meta description, and image alt text. Language-pattern checks (verb detection, claim analysis, quotable sentence filtering) are adjusted for RTL script so you don't get penalized for patterns the tool can't reliably read in those languages.

When a non-English page is detected, a language notice appears above the results explaining what was detected and how scoring was adapted.

If the analyzer detects that very little visible text was returned from a URL fetch, it will show a warning notice and stop the analysis. This prevents a misleading low score from being displayed when the real page content was never actually received.

The warning reads: "This page may not have been fully captured. Some pages are JavaScript-rendered and load their content dynamically, which means they can't be fully fetched by URL alone."

A related but different scenario: the page HTML is fetched successfully, but the content extractor finds very little readable text while the raw page clearly has visible structure (buttons, headings, pricing blocks). In this case the analysis still runs, but a high-priority warning appears in the Action Plan:

"Content Extraction Mismatch — The page has visible structure, but the analyzer could not extract meaningful AI-readable content from it."

This typically affects complex interactive components — tab-based pricing tables, JavaScript-toggled accordions, or heavily componentised pages where key content is locked inside dynamic UI. The fix is the same: use the HTML paste method to give the analyzer your browser's fully rendered view, or expose key content in semantic HTML and structured data.

Use the HTML paste method — your browser has already run the JavaScript and rendered the full page, so copying the source gives the analyzer complete content.

Check the Action Plan — it's usually one or two high-impact gaps pulling the score down significantly. The most common culprits are missing FAQ schema, no author attribution on blog content, sentences that are too long for AI extraction, and claims without supporting evidence.

There are two reasons a check may not appear or may be suppressed:

Page type: The tool detected your page type and suppressed checks that aren't relevant. For example, a pricing page won't be penalised for missing an author byline. N/A checks don't count for or against your score.

Thin pages: Some checks are suppressed on pages with very little content — FAQ and Statistics checks require at least 500 words, and the Quotable Sentences check requires at least 400. On shorter pages there simply isn't enough material to assess these meaningfully, so the check is skipped rather than producing a misleading score.

Use the HTML paste method instead. It works on every page regardless of how the site is configured — including sites that block external crawlers. If the analyzer detected a JavaScript-rendered page, it will show a warning notice and pre-fill the URL field for you automatically — just paste the HTML and click Analyze HTML. See the JavaScript-rendered pages section for full details.

No — it significantly improves the likelihood. Actual citation also depends on query match (whether someone asks something relevant to your page), competition (what other pages exist on the same topic), and freshness. The score reflects what you can control on the page itself.

After any significant content update, after adding schema markup, or after publishing new supporting content that links to the page. For high-priority pages, a monthly check is reasonable.

Yes — many GEO improvements directly improve traditional SEO signals too. Better structure, stronger E-E-A-T signals, clearer headings, and added FAQ schema all help with Google as well as AI engines.

Yes — the downloaded report is a standalone HTML file that works in any browser with no internet connection required. You can share it directly, attach it to emails, or save it to a shared drive.