AvocadoScore

Your PageSpeed score + how Claude, GPT and Gemini see you.

A ranked fix list your engineer can ship — every fix, every AI bot audited, completely free. No paywall, no signup, no catch.

Free, full report  ·  No signup  ·  Every fix unlocked

What you get

A real report — not a marketing screenshot.

Below is a real audit of vercel.com 10 pages. Click anything to open the full report.

Verdict

Almost ripe

Mobile perf

71/100

SEO ready

96/100

AI ready

89/100

OG image

100/100

image/png

llms.txt

100/100

24 links · 9 sections

Schema coverage

4/10

BreadcrumbList · Organization · TechArticle

Open the full report

Your next customer just asked ChatGPT, Claude, or Perplexity to recommend a tool in your category. If AI can't see your site, it recommended someone else.

Lighthouse can’t tell you what ChatGPT thinks of your site. We can.

CapabilityPageSpeedLighthouseAhrefsAvocadoScore
Core Web Vitals (real + lab)
Site-wide scan in one click
SEO checklist with fix list
AI bot policies (GPTBot, ClaudeBot…)
llms.txt audit
Plain-English fix list w/ effort estimates
Site-wide report at one shareable URL

Deeper comparisons: vs Lighthouse · vs PageSpeed · vs Ahrefs · vs Semrush

What AI readiness actually means.

No black box. 18 specific checks across three pillars, so an AI agent can find, parse and credit your site.

AccessCan AI crawlers fetch your site at all?
  • Answer-engine crawlers
    Whether robots.txt blocks the bots that cite you — OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-User. Training opt-outs (GPTBot, CCBot…) are reported separately and never penalized.
  • CDN / firewall gate
    We fetch your homepage AS an AI bot and diff it against a normal browser — robots.txt is what you say, the CDN is what you do (Cloudflare blocks AI crawlers by default).
  • Search-index crawlers
    ChatGPT search and Copilot ride Bing’s index; AI Overviews ride Google’s. Blocking Bingbot removes you from AI answers too.
  • Snippet directives
    nosnippet and max-snippet quietly cap what answer engines may quote from you.
ComprehensionOnce fetched, can they parse it?
  • /llms.txt + /llms-full.txt
    The emerging standard letting AI agents index your site cleanly.
  • Validated structured data
    Not just “JSON-LD present” — we check required properties per type: a Product without offers can’t be quoted with a price.
  • Server-rendered content
    AI crawlers don’t execute JavaScript. We detect empty JS app shells outright.
  • Heading outline & semantic HTML
    LLM retrieval is chunk-based — a clean H1→H2→H3 hierarchy is what gets sections cited.
  • Feeds, sitemap freshness, consent walls
    RSS/Atom ingestion channels, <lastmod> recrawl signals, and whether crawlers see your content or your cookie banner.
TrustCan they attribute and date it?
  • Author / Organization schema
    Provenance signals AI assistants increasingly weight for trust.
  • Content dates
    datePublished / dateModified — without them, answer engines can’t judge freshness.

+ standard Core Web Vitals (6 metrics) and a 12-point SEO checklist included in every audit.

Want the AI side checked in 5 seconds, no scan needed? Try the free readiness checker →

Even top SaaS sites leave AI-readiness points on the table.

Real reports we ran below. Click any to see the full audit, the fix list and the bots blocking the site from showing up in answers.

Curious how your site scores? Scan it free →

The AI Readiness Index

Who’s actually ready for ChatGPT?

50 well-known sites, ranked by AI readiness — the same free scan, run on names you know. Updated daily.

See all 50, ranked → or find out where you’d land →

Three steps from URL to fix list.

01

Paste your URL

Just the domain. No signup, no install, no setup.

02

36 checks, one pass

Core Web Vitals · 12-point SEO scorecard · 18 AI-readiness checks across 12 crawlers, three pillars: Access, Comprehension, Trust.

03

Get your report

A shareable URL with per-page deep dives, fix list, and effort estimates.

Free. The whole report. No catch.

Every scan is the full audit now — every fix, every page we scan, nothing locked. We monetize the tools that keep it working for you over time, not the report itself.

Full audit

Up to 10 pages, every fix unlocked, forever free.

$0

  • Up to 10 pages — picked from your sitemap
  • Full SEO + AI Readiness scorecards
  • Every fix: copy-paste snippet, steps, AI prompt
  • Public shareable URL + embeddable badge
  • No signup, no card, no expiry
Recommended

Monitor

Keep it fixed — we watch the site so you don't have to.

$9/mo

  • Monthly re-scan, automatically
  • Email alert the moment a score slips
  • Score-over-time trend on your report
  • Same free report underneath — this just watches it
  • Cancel any time

Running an agency or a big portfolio of sites? Email us — we're building tools for exactly that.

Independent. No funding, no tracking, no VC roadmap.

Actually free — no card, no expiring preview, every fix unlocked. A real human answers every email — usually the same day. And we run the same audit on our own site.

Questions, answered.

Is this safe to scan? Will it slow my site?
We use Google's public PageSpeed Insights API, which hits your site exactly like Googlebot — a handful of normal page loads. No special permissions, no user data captured.
Is it actually free? What's the catch?
No catch. The full report — every fix, every page we scan, the copy-paste snippets and the AI prompt — is free, no card, no signup, no expiring preview. We keep it cheap to run with a daily cache and a generous-but-capped page limit, and we make money from Monitor, an optional $9/mo subscription for anyone who wants automatic monthly re-scans and regression alerts. Skip Monitor entirely and the audit stays free forever.
What if our engineers push back on the recommendations?
Every finding cites the underlying metric (LCP, CLS, INP, schema field, AI-bot directive) and links to the relevant spec. The action plan is ranked by effort × impact, so eng owns the trade-off — we don't prescribe sprints. If a finding is genuinely wrong for your stack, email us with the scan ID and we'll fix it.
How is this different from Lighthouse or PageSpeed Insights?
Lighthouse and PSI are tools — they give you raw numbers, one page at a time. AvocadoScore turns those numbers into a site-wide report with a 12-point SEO scorecard, an 18-point AI-readiness check across every major AI crawler (GPTBot, ClaudeBot, PerplexityBot and more — including whether your CDN blocks them) and the new llms.txt standard, and a plain-English fix list with effort estimates — free. The kind of report Lighthouse can't write.
How long does a scan take, and how many pages?
Under a minute for up to 10 pages, picked from your sitemap. That covers the homepage plus the pages that usually matter most (pricing, docs, key templates). Monitor subscribers get a deeper monthly scan — up to 100 pages, median-of-3 PageSpeed runs per page for stable numbers.
Can I scan a staging site or one behind a login wall?
Our scans run from Google's PageSpeed Insights infrastructure, which can only see what's publicly reachable. If your staging URL is on the public internet (even with an obscure subdomain), it scans fine. If it's behind basic auth, an IP allowlist or a login wall, we can't reach it — email us and we'll figure out whether a temporary preview link works.
Can I re-scan after we ship the fixes?
Yes — just run it again, free, any time; recent scans of the same domain are served from cache for a day so repeat checks are instant. If you want it to happen automatically, Monitor ($9/mo) re-scans monthly on its own and emails you the moment a score slips, with a trend view of what changed since last time.
Who can see my report? Where does it live?
Every report sits at an unguessable UUID URL like /r/<random>. Free reports are public by link — anyone with the URL can view, nothing search-indexed, marked `noindex` and never listed in a sitemap. Monitor reports (tied to your subscription) are gated: a magic-link sets an access cookie scoped to that report, and forwarding the email extends the same access to whoever you send it to.

Ship the audit your engineer will actually read.

Free, every fix unlocked. No card, no signup, no catch.

Not ready to commit? Try these first.