pigfox.ai: Can Answer Engines Read Your Website?
pigfox.ai is a free check by Pigfox: whether ChatGPT, Claude, Perplexity and other answer engines' crawlers may read your site, plus llms.txt, structured data, title and description, scored out of 100.
More people now ask ChatGPT, Claude or Perplexity a question and read the answer, instead of scanning ten blue links. Those answers are built from pages the engines’ crawlers were allowed to read. If your robots.txt turns those crawlers away, or your pages give them little to work with, your business is simply not part of the answer.
pigfox.ai is a free check, built by Pigfox, that tells you where you stand. Type a domain, and a few seconds later you get a score out of 100 and a short list of what to fix.
What it checks
Crawler access in robots.txt. pigfox.ai reads your robots.txt the way the crawlers themselves
read it (the RFC 9309 rules: a crawler follows the group that names it, or else the * group, and
the longest matching rule wins). It reports the answer for your homepage for each crawler by name:
OpenAI’s GPTBot, OAI-SearchBot and ChatGPT-User; Anthropic’s ClaudeBot, Claude-SearchBot and
Claude-User; PerplexityBot and Perplexity-User; Google-Extended; Applebot-Extended; CCBot;
Bytespider; Meta-ExternalAgent; Amazonbot; DuckAssistBot; and the * default. For each one it says
what that crawler is for, because they are not all the same: some fetch pages for search results, some
open a page because a person asked, and some collect training data.
llms.txt. Whether the site publishes one, and whether it starts with a # heading as the format
expects.
The homepage itself. The HTTP status and final address, the title and its length (30 to 60 characters reads best), the meta description, the canonical link, the language, any noindex in meta robots or X-Robots-Tag, whether there is an h1, and every JSON-LD block with its Schema.org types.
A score, and what to do about it
Each check is worth a set number of points, and the report shows the total out of 100. More useful than the number is the list under “What to fix”: each item says what is wrong, why it matters, and how many points fixing it is worth.
A worked example: triremes.com
We ran pigfox.ai on our own flagship, triremes.com, on 8 November 2026. It scored 96 out of 100, rated Ready.
Every search and answer crawler on the list is allowed, among them OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Amazonbot and DuckAssistBot, along with GPTBot, ClaudeBot and Google-Extended. The missing points come from three crawlers that triremes.com’s robots.txt blocks by name, each worth 1.4 points:
- Meta-ExternalAgent, which collects pages for training Meta’s models;
- CCBot, which builds the open Common Crawl dataset that many models train on;
- Bytespider, which collects pages for ByteDance’s models.
None of the three answers anyone’s question; they collect training data. pigfox.ai scores what each crawler is allowed to do, but whether you want a training crawler on your site is your call, and the report says which crawler is which so you can make it. pigfox.com itself scores 100 out of 100.
Free, and private by design
pigfox.ai is free, with no account and no sign-up. In the words of its privacy page: no cookies, no accounts, no ads, and visits are counted without cookies and without storing your IP address. To tell visitors apart within a day, the server hashes your IP address and browser with a random salt that changes every day and deletes yesterday’s, so a visit today can’t be linked to one tomorrow; the IP address itself is never written to the database or a log, and the counts are deleted after about 13 months. There is no third-party analytics and no third-party script, and the pages load nothing from other servers. A check’s result is kept in memory for 10 minutes, so checking the same site twice does not fetch it twice.
It is polite to the sites it checks, too. A check fetches four things: the homepage (at most the
first 1 MB), robots.txt, llms.txt, and a HEAD request for llms-full.txt. It identifies itself as
PigfoxAIReadiness/1.0, follows no links, and connects only to public addresses.
pigfox.ai is built by Pigfox in Go.
Where Triremes comes in
pigfox.ai checks one site, the one you type, when you ask. Triremes, our flagship, measures search and answer-engine readiness across business websites at scale: as of 8 November 2026 it has measured and dated more than 550,000 of them. It also collects the problems people keep reporting with the software they use. Businesses with poor SEO or AEO scores are leads: you can buy the lead, the idea report or the website, and pitch or build the fix.
If pigfox.ai’s report shows your own site needs work, the list under “What to fix” is the place to start. If you build or market websites for others, Triremes is where to find the sites that need you.