Run a Beacon scan on your site.
We crawl your site against four AI search bots and grade how visible your content is to ChatGPT, Claude, Perplexity, and Google’s AI mode.
Schema & structured data. Crawlability and SSR. Content structure and answer-first formatting. AI crawler directives (llms.txt, GPTBot, ClaudeBot, PerplexityBot, Google-Extended). It’s the layer beyond SEO that decides whether AI search engines cite you.
Below 80?
Beacon Improve lifts you above it. Two weeks, guaranteed.
- ✓Structured data (JSON-LD): Article, Product, FAQ, Org, Breadcrumb
- ✓AI-crawler directives: llms.txt + robots.txt for GPTBot, ClaudeBot, PerplexityBot
- ✓Crawlability fixes: SSR hydration, canonical tags, sitemap completeness
- ✓Monitoring: Beacon tracker installed, 30-day follow-up re-audit included
Eligibility: Beacon score must be <80 to guarantee +20 lift.
Almost Impossible Agency, scanned in 30 seconds.
The output: a four-bot composite score, per-check breakdown of what each AI search engine actually reads, and the prioritized fix list. No signup, no email, no nag.
Honest answers to the obvious ones.
What's actually being scored?
A composite of: robots.txt accessibility for GPTBot, ClaudeBot, PerplexityBot, Google-Extended; structured data validity; llms.txt presence; semantic readability for non-rendering bots; mention density. Each bot's score is independent; the composite is a weighted average.
How accurate is it?
Beacon mirrors what each bot can actually read on its first pass. We calibrate against a corpus of ~50 known-citable sites and grade against their AI-Overview / ChatGPT citation rate. Last calibration: April 2026. Composite scoring is exact; per-bot is ±5 points.
How is this different from Lighthouse / Yoast / Schema Pro?
Those score against Googlebot's rendered DOM. Beacon scores against GPTBot, ClaudeBot, PerplexityBot, and Google-Extended which see different content. A site that scores 95 on Lighthouse can score 30 on Beacon.
Is anything stored or shared?
Each scan is a single HTTP fetch per bot profile from a Vercel edge function. We log URL + score + timestamp for accuracy calibration. No email captured. No retargeting.
What happens after I see my score?
Above 80, the report names the cheap wins. Below 80, you get Beacon Improve as an option (two-week sprint, +20 points or we keep working). Either way, the report is yours to keep.