A free check of how your site looks to Google and to AI assistants like ChatGPT, Claude and Perplexity. Plain-English findings, each with a fix prompt you paste straight back into your AI tool.
We ask Claude, with web search on, five questions a buyer asks about your category, never naming you, and count how often it names you, cites your site, and names your competitors instead. Then we search for you on the directories buyers use. Two a day each; the answers are a snapshot, not a ranking.
We read your home page and up to 20 more from our server, as vibecheck-visibility, and follow your robots.txt. No sign-up. Results are kept for 24 hours so a busy site isn't crawled again and again; if you fill in the AI panel, what you typed there is kept with them and sent to Anthropic inside the questions. Privacy.
Whether your words are in the HTML or only drawn by JavaScript (the blank-page problem most Lovable, Bolt and Vite apps ship with), robots.txt, the sitemap and its dates.
llms.txt, which AI crawlers you let in, structured data for who you are and what you charge, and machine-readable files.
Titles, descriptions, one clear H1, link previews, sections stuck behind #fragments, alt text.
www against bare, HTTP against HTTPS, redirect chains, canonical tags, links that go through redirects.
A way to get in touch, a privacy policy, who's behind it, and what it costs.
Every finding says why it matters, gives a prompt for your AI tool, and tells you how to see the fix worked.
Why a fresh Lovable or Bolt app is a blank page to most crawlers, and the fix that keeps your app as it is.
A plain-text summary for AI assistants. What goes in it, what it does and doesn't do, and a template.
Five reasons ChatGPT, Claude and Perplexity never mention a new app, and what to fix first.
Training crawlers and search crawlers are different bots. Which to keep out, which to let in.
Requests from this page come from our server with the user agent
vibecheck-visibility/1.0 (+https://psychosecurity.io/visibility). Someone typed your address into
this page, or ran the command-line tool against it.
A check reads robots.txt, sitemap.xml and llms.txt, your home page and up to 20 more pages, the other form of
your domain (www or bare), the http:// address, your link-preview image and the sizes of up to 12 images. At most
six requests at a time, and one check per site per minute however many people ask. It never submits forms or
follows links off your site, and it obeys robots.txt rules for vibecheck-visibility. To keep it out:
User-agent: vibecheck-visibility Disallow: /
The same check, from your own machine, with no limits. Python 3.9+, no dependencies. It exits 1 on any high or critical finding, so it can run in CI.
python3 -m vibecheck --visibility https://myapp.com --markdown report.md python3 -m vibecheck --visibility https://myapp.com --json