The Uncomfortable Truth About llms.txt Generators (And Why Most Are Lying to You)
The Uncomfortable Truth About llms.txt Generators (And Why Most Are Lying to You)
December 19, 2025 | By Jamie Watters
I spent the last month testing every llms.txt generator I could find. What I discovered made me angry - and I think you should be angry too.
The Promise vs The Reality
When AI search became mainstream in 2024, a new standard emerged: llms.txt. A simple text file that tells ChatGPT, Claude, and Perplexity which pages on your website are worth citing.
Simple, right?
That's what the free tools want you to believe. "Enter your URL, get your llms.txt file in seconds!" They show you impressive numbers: "We found 133 pages!" "1,500+ URLs discovered!"
And technically, they're not lying. They did find those pages.
But here's what they're not telling you: Finding pages is the easy part. It's what you do with them that matters.
The Junk Problem
Last week, I ran the same website through four different llms.txt generators. FreecalcHub.com - a calculator site with about 150 pages.
Results:
| Tool | Pages Found |
|---|---|
| LLM.txt Mastery (mine) | 147 |
| Keploy (Free) | 133 |
| SiteSpeakAI | 21 |
| Writesonic | 53 |
Look at those numbers. Keploy found almost as many pages as we did. On quantity alone, you might think "Free tool works fine, why pay?"
Then I looked at what they actually produced.
Keploy's output: A raw dump of 133 URLs. The homepage. The privacy policy. The terms of service. The cookie consent page. The contact form. All mixed together with the actual valuable content.
No prioritization. No quality indicators. No guidance whatsoever on which pages matter.
I called it what it is: junk.
Why This Matters (And Why It's Hard to Explain)
Here's my marketing problem: Most people don't understand why quality scoring matters. They see "133 pages found" and think the job is done.
It's not.
When Claude or ChatGPT crawls your llms.txt file, they don't have infinite context. They can't read everything. They have to choose which pages to prioritize.
If your llms.txt file treats your privacy policy the same as your flagship product page, you're telling AI systems: "These are equally important."
They're not.
Quality scoring solves this. Every page we analyze gets a 1-10 score based on:
- Content depth and uniqueness
- Technical relevance
- Information architecture
- AI citation potential
Pages scoring 8-10 get included. Pages scoring 3-4 don't. Your privacy policy probably scores a 2. Your detailed product comparison guide? That's an 8 or 9.
The difference between being found and being cited is knowing which pages matter.
What We Built This Week
Frustrated by the gap between what users need and what they're being sold, we shipped two major features:
1. Platform-Specific Deployment Guides
After generating your llms.txt file, we now detect what platform you're using (WordPress, Netlify, Shopify, Vercel, React, Next.js... 17 platforms total) and show you exactly how to deploy it.
Not generic "upload to your root directory" nonsense. Actual step-by-step instructions for YOUR stack.
WordPress user? Here's how to use cPanel. Netlify user? Here's the exact folder structure. React app? Here's where it goes in your public folder.
Verification checklist included.
2. API Tiering for Partners
Our API now supports tiered access. Partners like AImpactScanner can pass their user's subscription tier and get appropriate service levels - including JavaScript rendering for Scale tier users.
This matters because 60-80% of modern websites use JavaScript frameworks (React, Vue, Angular). Basic crawlers can't see their content. They return empty pages or broken fragments.
We actually render the JavaScript. We see what AI systems will see.
The Wheat and the Chaff
Here's my honest assessment of the llms.txt landscape:
Free tools (Keploy, etc.): They work for page discovery. If you just want a list of URLs, they're fine. But they're lead magnets, not serious tools. No quality scoring. No framework detection. No ongoing development.
Enterprise tools ($879+/year): They exist, but they're priced for Fortune 500 companies. Most solopreneurs and small businesses can't justify that cost.
Us ($0-$20/month): We're the only dedicated llms.txt platform with quality scoring, framework detection, JavaScript rendering, and active development (47+ updates shipped in 2025 alone).
The gap between "free and junk" and "$879/year" is exactly where we live.
Why It's Hard to Get the Truth Out
I'll be honest: Marketing this is frustrating.
When a competitor shows "1,500 pages found!" people get excited. When I explain "Yes, but 1,200 of those are junk and will actively hurt your AI visibility," their eyes glaze over.
Quality is invisible until it matters.
The businesses that get it are the ones who've already experienced the problem. They ran a free generator, uploaded the file, and... nothing happened. Or worse - AI started citing their privacy policy instead of their actual products.
For everyone else, we're fighting an uphill battle against shiny numbers.
What You Can Actually Do
If you're evaluating llms.txt generators, here's my honest advice:
-
Test them on your own site. Don't trust landing page claims.
-
Look at the actual output. Is it a raw URL dump or a structured file with priorities?
-
Check for quality indicators. Are high-value pages distinguished from boilerplate?
-
Verify framework support. If you're using React/Vue/Angular, most tools will fail silently.
-
Check robots.txt compatibility. This is our unique feature - we detect when your robots.txt conflicts with your llms.txt. No one else does this.
Free tools are fine for discovery. But if you want AI systems to actually cite your content - not just find it - you need quality intelligence.
That's the difference between a URL dump and a strategic asset.
The Numbers That Matter
After today's deployment:
- 17 platforms with specific deployment instructions
- 15+ JavaScript frameworks detected and rendered
- 4 subscription tiers with API access
- 47+ updates shipped in 2025
- $0 to start (Starter tier is free, forever)
We're building the tool I wished existed when I started.
Jamie Watters is the founder of LLM.txt Mastery, the only dedicated llms.txt platform with AI quality scoring. He builds tools for solopreneurs and writes about the honest reality of competing with well-funded startups and "free" lead magnets.
Try it free: llmtxtmastery.com
Tags: #buildinpublic #llmstxt #ai #seo #solofounder