Cite us freely.
Cabbge measures how often AI engines (ChatGPT, Gemini, Perplexity, Claude, Grok) recommend brands when buyers ask. Reporters and researchers can cite anything on this page freely.
The pitch.
Cabbge is a self-serve GEO (Generative Engine Optimization) platform measuring how often AI engines recommend brands when real buyers ask. It runs real prompts on ChatGPT, Gemini, Perplexity, Claude, and Grok, scores per-engine visibility, audits AI-readiness signals on a brand's site, and audits off-domain coverage on Wikipedia / Wikidata / G2 / Trustpilot / Reddit. It also tracks every public mention of a brand across Reddit, Hacker News, YouTube, and X, and ships schema, FAQs, articles, and a personalized cold-outreach kit to close the gaps. Built for indie founders, bootstrapped SaaS, ecom and Shopify operators, and small marketing teams. Self-serve only, $49 Starter and $199 Pro, no demo calls.
The numbers.
No brands graded yet. Check back after seed cron runs.
Reporter angles.
- /01THE WIKIPEDIA TAXBrands with a Wikipedia article and Wikidata entity score on average 18 points higher than brands without — across all 5 engines. Concrete number to anchor an 'AI search rewards old institutional content' piece.
- /02THE JS-RENDER TRAPCategories where the leading SPA-built brands disappear from ChatGPT entirely (which uses raw HTML). Useful for a 'why your single-page app is invisible to AI' story.
- /03ENGINE DISAGREEMENTThe most contested category — brands ChatGPT recommends that Gemini doesn't, or vice versa. Drives the 'there is no one AI search' angle.
- /04REDDIT AS THE NEW RANKING FACTORBrands with high Reddit mention counts (visible in our tracker) rank disproportionately well on Perplexity. Data for 'Reddit is the source AI engines trust most.'
The pipeline.
Full methodology lives at /methodology. Short version: real ChatGPT, Gemini, Perplexity, Claude, and Grok APIs (no scraping); 8 prompts on the free public grader, up to 40 on paid scans; mention detection includes a deterministic verification pass to suppress LLM hallucinated positives; scoring blends per-engine share-of-answer with a 13-point AI-readiness audit and a 5-source off-domain coverage check.
How to attribute.
Quote any score on a public /visibility/[domain] page. Each is a permanent shareable URL with the scan date, engine breakdown, and methodology link. We refresh public grades every 7 days; cite the "scanned" date on the page.
Suggested attribution: "according to Cabbge's AI visibility grader" or "cabbge.com/visibility/[brand] (May 2026 scan)".
Quotes available.
Subject line "press" gets fast routing. Available for quotes on AI search, GEO, AI visibility for SMBs, and the broader market shift from Google to generative answers.
HI@CABBGE.COM →