AI Visibility Checker Version 2: Free Tool Now Reads Your Files Like ChatGPT Does

The free AI Visibility Checker at 365i just got a proper upgrade. Not a cosmetic tweak. A fundamentally different tool.

Version 1 was a glorified file detector. Does llms.txt exist? Tick. Does it return a 200 status? Tick. Done. But after personally testing over 100 websites in the past fortnight, I kept hitting the same problem: having files isn't the issue. Having broken files is.

Version 2 changes the approach entirely. Instead of checking if files exist, it reads them the way ChatGPT, Claude, and Perplexity actually would. AI-powered analysis of content, structure, and consistency. The results have been eye-opening.

What Changed in Version 2

Version 1 was binary. File exists? Yes or no. Returns 200 status? Yes or no. Helpful for spotting completely missing files, useless for finding broken ones.

Version 2 uses actual AI systems (Claude and GPT) to read your files the same way ChatGPT would when answering questions about your business. Here's what it analyses:

• Content quality: Does the file contain proper canonical identity blocks?
• Structure: Is it formatted correctly for AI parsing?
• Consistency: Do service descriptions match across all files?
• Completeness: Are required fields actually present?

Instead of ticking boxes, you get actual feedback. "Your llms.txt has inconsistent service descriptions" or "Your canonical identity block is missing the location field." Diagnostic information you can act on.

And the clever bit: it shows you how AI interprets your current files. You can see exactly what ChatGPT would understand about your business based on what you've written.

The Free Template Library

This bit surprised even me. Version 2 now includes free downloads of all ten professional AI identity file templates. These are the same templates used in the £295 paid service, just without the custom research and implementation.

Why give them away? Because showing businesses what "100 out of 100" looks like is more valuable than guarding templates. You download them, see proper structure, understand the canonical identity format, and decide whether to DIY or get professional help.

The templates cover all ten files in the AI discovery file specification: llms.txt (primary AI guidance), llm.txt (compatibility variant), llms.html (human-readable version with Schema.org), ai.txt (usage and intent signals), ai.json (machine-readable interaction guidance), identity.json (canonical business identity), brand.txt (brand representation rules), faq-ai.txt (structured Q&A), developer-ai.txt (technical context), and robots-ai.txt (AI-specific directives). Our complete guide to AI discovery files explains what each one does in detail.

The Industry Context

AI-referred traffic jumped 527% between January and May 2025 across 400+ websites that were analysed. Not a typo. Five hundred and twenty-seven percent.

ChatGPT, Claude, and Perplexity aren't answering the odd question anymore. They're replacing Google for millions of queries daily. When someone asks "recommend a good WordPress host in the UK" or "find me a plumber in Kettering," they get direct answers. If AI doesn't know about you, or gets your business wrong, you're invisible.

"It's 2025 and most content is still written for humans instead of LLMs. 99.9% of attention is about to be LLM attention, not human attention."

Jeremy Howard, founder of FastAI, Publii Complete Guide to llms.txt

When I first read that, my reaction was scepticism. 99.9% felt like hyperbole. But running a hosting company since 2001, I've seen these shifts before. Mobile traffic, voice search, social media as a discovery channel. Each time the early adopters had an outsized advantage. The 527% growth stat suggests we're in exactly that kind of shift right now.

Wix Studio's AI Search Lab reported in November 2025 that 64% of technology executives plan to deploy agentic AI within 24 months. That's AI systems actively browsing websites on behalf of users, not just answering from training data. When an AI agent browses your site, it needs to understand what you do quickly. Your homepage might be 300KB of HTML. A structured llms.txt file? 2KB. The tokenisation cost difference is enormous at scale.

BuiltWith tracking shows over 844,000 websites had implemented llms.txt by October 2025. Anthropic, Cloudflare, Stripe, Perplexity, Zapier, Cursor, Yoast, and DataForSEO among them. Not a niche list.

Real-World Findings from 100+ Sites

Over two weeks, I personally tested over 100 websites with Version 2. Not automated scraping, actual manual analysis. Four patterns kept showing up.

Most sites with files still scored poorly. Having llms.txt wasn't enough. One business had the file but with formatting that broke AI parsing. Another had all ten files but with contradictory information across them.

Canonical identity blocks were frequently missing or incomplete. AI needs consistent identity information: business name, what you do, where you operate, who you serve. Many files had some of this scattered in ways AI couldn't reliably extract.

Service descriptions were vague. "We provide web solutions" tells AI nothing useful. "WordPress hosting with 100% SSD storage, free CDN, and UK-based support" gives AI concrete facts to recommend you for. If you're wondering why ChatGPT can't find your website, vague descriptions are often the culprit.

Multiple files telling different versions of the truth. This was the most common problem. llms.txt says one thing, ai.json says something slightly different, faq-ai.txt has a third version. AI doesn't know which to trust, so it trusts none of them.

The sites that scored well had consistent, factual information across all files with proper canonical identity structure. Not complicated, just consistent.

The Controversy Nobody's Discussing

Not everyone agrees this file format matters. And I think the honest disagreement is worth covering.

"llms.txt has different purposes [from robots.txt]. robots.txt is generally used to let automated tools know what access to a site is considered acceptable... llms.txt information will often be used on demand when a user explicitly requests information."

llmstxt.org, Official Specification

I've been turning this distinction over for months now. The comparison to robots.txt history is what convinced me the format has legs. We had years of "do search engines even respect robots.txt?" debates in the early 2000s before it became a de facto standard. The trajectory feels familiar.

Google's John Mueller dismissed llms.txt in July 2025, comparing it to the discredited keywords meta tag. Fair point. If AI systems need business information, they can extract it from your website content.

But then in December 2025, Google quietly added llms.txt to their Search Central documentation. When SEO professionals noticed, Mueller responded with "Hmmn :-/" and the file disappeared from Google's docs within hours. Make of that what you will.

The research is mixed too. SE Ranking analysed nearly 300,000 domains and found that having llms.txt "doesn't impact how AI systems see or cite your content today." Search Engine Land implemented it and monitored server logs from August to October 2025. Zero visits from GPTbot, PerplexityBot, or ClaudeBot.

So why bother? Two reasons. First, tokenisation efficiency. When agentic AI browses thousands of sites, a 2KB text file that can be parsed 286 times for the same cost as a 300KB homepage matters. Second, the format addresses a real problem that will only grow as AI agents become more common. Think of it like schema markup in 2011: initially niche, dismissed by some, adopted by major tech companies, eventually essential.

What You Should Actually Do

1. Run the free checker. Head to the AI Visibility Checker. Takes about 30 seconds. You'll get a score out of 100 plus AI-generated analysis of each file.

2. Read the AI-generated feedback. Don't just look at the score. Read what AI extracted from your files. Is it accurate? Complete? Does it represent your business correctly?

3. Download the free templates. See what proper structure looks like. Compare to your existing files. Understand why consistency matters more than word count.

4. Decide your approach. Two honest options:

• DIY: Budget 5-10 hours. Use the templates, follow our step-by-step guide to creating AI files, test thoroughly. Completely doable if you're technical.
• Professional service: £295, delivered in 48 hours. We handle research, implementation, consistency checking, and validation for all ten files.

Don't frame this as urgent. Frame it as preparation for an emerging standard with real technical merit behind it. Even if AI platforms never officially adopt llms.txt, having consistent, structured information about your business in machine-readable formats isn't a waste of time. Worst case, you've created clear documentation of what your business does.

844,000 websites including Anthropic, Stripe, and Cloudflare didn't implement this for no reason. Whether the payoff comes in six months or two years, the businesses building their AI identity now won't regret being early.

Frequently Asked Questions