If AI crawlers look, be worth reading.
llms.txt Studio writes a spec-compliant llms.txt from your own pages, grades the one you already have against the llmstxt.org format, and, if you subscribe, tells you when the file breaks and which crawler user-agents fetched it.
It does not promise traffic, rankings or citations. Nobody can. Whether an assistant reads the file is the assistant's choice; our job is that the file is correct, current and honest when it does.
A file, done properly.
- Generator: robots.txt → sitemap → up to 60 pages, ranked by family (docs, API, product, blog, about). Titles and descriptions are your own meta and headings. Deterministic: the same site gives the same file.
- Validator: 14 rules over the llmstxt.org format, link liveness on a sample of 20, content-type and size checks, robots.txt crawler notes. Every finding has a rule id and a line number.
- Monitor: weekly presence, grade and change checks, drift against the live site, and a fetch log of which crawler user-agents requested the file.
Not a ranking lever.
- No traffic claims. llms.txt is a proposal, not a standard any assistant has committed to. We report what the file says and who fetched it, not what it earned you.
- No generated prose. The summary and descriptions come from your pages. If a page has no description, the file says so and you write one.
- No stealth crawling. Our fetcher announces itself, obeys robots.txt, stays under 2 requests per second and never follows links off your domain. Details.
Our parent brand's file, unedited.
This is the generator's output for echorank360.com: 60 links across Docs, Product, Blog, About and Optional, graded A by the validator, and deployed as-is at echorank360.com/llms.txt. The first customer is the parent company.
# Echorank360 > When customers ask ChatGPT, Google AI or Perplexity who to hire, is your business the answer? Measure your AI Visibility Score, track recommendations daily, and get a prioritized roadmap to become the business AI recommends. ## Docs - [Echorank Glossary — search, ranking & AI visibility terms](https://echorank360.com/en/glossary): The Echorank Glossary: definitions and concepts for search, ranking, AI visibility, and digital discovery. - [Echorank User Guide](https://echorank360.com/en/guide): By the end of this guide you will have connected your Google Business Profile, imported your review history, launched your first review campaign, set up… - [AI visibility, from zero to tracked](https://echorank360.com/en/guides/ai-visibility): What ChatGPT, Perplexity and Gemini answer when your market comes up — how to measure it, track it and improve it, step by step. - [Audit your website — what crawlers and AI actually see](https://echorank360.com/en/guides/audit-your-website): Run the free audit, read the itemised checks, and fix in the order that moves the score: crawlability first, structured data second, content last. - [Getting started with Echorank — setup guide](https://echorank360.com/en/guides/getting-started): From an empty account to a working dashboard in about 15 minutes: create your account, connect Google Business Profile, run your first audit and import your… - [Keyword research — from customer language to a tracked list](https://echorank360.com/en/guides/keyword-research): Start from what customers actually say, expand in Keywords Explorer within your plan's search quota, filter by intent, and assign one primary keyword per page. - [Track rankings — a stable basket and a monthly read](https://echorank360.com/en/guides/track-rankings): Pick ten to twenty-five terms, add them to Rank Tracker, read trends over four weeks rather than four days, and connect Search Console for the queries you… - [Learn reputation & AI visibility — the Echorank Knowledge Hub](https://echorank360.com/en/learn): A free ten-chapter course, five deep-dive guides and a glossary: audit your baseline, import your review history, generate and answer reviews, and get……
Fourteen rules, arithmetic shown.
Start at 100, subtract per rule hit up to a cap, letter by threshold. A missing file or an HTML page served in its place is fatal. Info rules are reported but never scored: blocking crawlers is a decision, not a defect.
| Rule | Severity | What it checks | Score |
|---|---|---|---|
| E1 | error | No file at /llms.txt | score = 0 |
| E2 | error | Not served as text, or larger than 1 MB | −40 each, max −40 |
| E3 | error | No H1 title | −25 each, max −25 |
| E4 | error | Malformed line inside a section | −5 each, max −25 |
| E5 | error | HTML page served instead of the file | score = 0 |
| W1 | warn | No blockquote summary | −10 each, max −10 |
| W2 | warn | No H2 sections | −20 each, max −20 |
| W3 | warn | Dead link | −4 each, max −24 |
| W4 | warn | Link points off-domain | −2 each, max −10 |
| W5 | warn | More than 200 links | −5 each, max −5 |
| W6 | warn | Duplicate URL | −2 each, max −10 |
| I1 | info | robots.txt blocks major AI crawlers | 0 (info) |
| I2 | info | llms-full.txt presence | 0 (info) |
| I3 | info | File age (Last-Modified) | 0 (info) |
Know when it breaks. Know when they look.
A weekly check that the file is still there, still valid, still matches your site. Email when the grade drops, a link dies, or the content changes without you. One alert per day at most.
Optionally, a canary line in your file points at a URL only we serve. When a crawler follows it, you see the timestamp and user-agent string. It proves the file was fetched and by which agent string. It cannot prove who is behind that string, and silence proves nothing.
See pricing- GPTBotOpenAI
- ClaudeBotAnthropic
- Claude-UserAnthropic
- Google-ExtendedGoogle
- PerplexityBotPerplexity
- BytespiderByteDance
- CCBotCommon Crawl
User-agent strings are self-declared and can be spoofed. We show them verbatim and label matches; we do not verify operator identity.