llms.txt: The New File Every Maritime Website Needs

In September 2024, a researcher named Jeremy Howard published a simple proposal. He suggested that websites should have a plain text file at their root domain, addressed directly to large language models, that explained what the site was about and pointed to the most important content. He called it llms.txt.

Within six months, hundreds of technology companies had adopted it. Developer documentation sites, SaaS platforms, and AI-native tools added the file almost immediately. The maritime sector, predictably, had almost none.

That gap is your opportunity. And the cost of closing it is one afternoon.

What llms.txt Is and Why It Was Created

To understand why llms.txt matters, you need to understand how large language models encounter and process websites. When a model like GPT-4, Claude, or Perplexity’s AI indexes the web for retrieval-augmented generation, it faces a fundamental problem: most websites are built for human visual processing, not machine comprehension. Navigation menus, footers, cookie banners, sidebar widgets, and JavaScript-rendered content all create noise that makes it harder for a model to identify what a site is actually about and which content is most valuable.

The robots.txt standard solved a version of this problem for traditional web crawlers by telling them which pages to index and which to skip. llms.txt solves the equivalent problem for AI: it tells language models what the site is, what it does, and where the most important content lives, in plain language that requires no interpretation.

The format is deliberately simple. It is a Markdown file. It opens with a heading that names the site, followed by a brief description, followed by sections that link to key pages with descriptions of what those pages contain. That is it. No schema, no special syntax, no technical implementation beyond creating a text file and placing it at the right URL.

The simplicity is strategic. A file that required technical expertise to create would not be widely adopted. A file that any website owner can create in twenty minutes has a real chance of becoming a standard, which is exactly what it is becoming.

How AI Crawlers Use llms.txt

When an AI system indexes your site for retrieval, it is making decisions about what to read, in what order, and how much weight to give each piece of content. Without llms.txt, these decisions are made by inference: the model looks at your sitemap, your navigation structure, your page titles and meta descriptions, and tries to construct a mental model of what your site offers and to whom.

This inference process works reasonably well for sites with clean architecture and clear content hierarchy. It works poorly for sites that have grown organically over years, have legacy pages that do not reflect current services, or have content buried in subdirectories that the crawler does not prioritise.

llms.txt short-circuits the inference process. Instead of the model having to guess what your most important pages are, you tell it directly. Instead of it having to infer what your company does from scattered content across multiple pages, you state it clearly in two sentences at the top of the file. Instead of it assigning equal crawl priority to your valuable service pages and your outdated blog posts from 2019, you give it a prioritised list.

The practical effect is that AI systems that encounter your llms.txt file build a more accurate and more complete model of what your site offers, which makes them more likely to cite you when they synthesise answers to relevant queries.

Perplexity has stated that it uses llms.txt files when available to improve retrieval accuracy. Several other AI search providers have confirmed similar behaviour. As AI search continues to grow, the sites with well-structured llms.txt files will consistently outperform those without in AI-generated responses.

What Maritime Companies Need to Include

A generic llms.txt file is better than no llms.txt file, but a well-constructed one that reflects the specific nature of your maritime business will perform significantly better. Here is what to include:

The opening description

Two to four sentences that describe your company with the precision of a Wikipedia article. Avoid marketing language. State what you do, who you serve, where you operate, and what makes your service distinct. An AI system reading this description should be able to answer the question “what is this company?” accurately without reading any other page on your site.

Bad: “We are a leading maritime services provider committed to excellence and innovation in the global shipping industry.”

Good: “MaritimeCo is a crew management and manning agency based in Manila, Philippines, specialising in the supply of certified deck and engine officers for bulk carrier and tanker operators in the Asia-Pacific region. The company manages the full employment lifecycle from certification verification through repatriation for approximately 400 seafarers annually.”

The second version gives an AI system everything it needs to know when deciding whether to cite your company in response to queries about crew management, manning agencies, or seafarer employment in Asia-Pacific.

Service sections with page links

For each major service or product area, create a section in llms.txt that names the service, provides two to three sentences of description, and links to the relevant page on your site. The description should include the specific vessel types, trade routes, certifications, or regulatory frameworks that your service relates to, because these are the terms that will appear in the queries your potential clients are asking.

Geographic coverage

Maritime is an inherently geographic industry. Port agencies, towage providers, ship chandlers, and marine surveyors all operate within defined geographic territories. Your llms.txt should make your service area explicit, because AI search is increasingly used for geographically specific queries: “ship chandlers in Rotterdam,” “marine surveyors in Singapore,” “crew management companies with Adriatic coverage.” If your territory is not in the file, you will not be in the answer.

Regulatory and certification context

Maritime operates under a complex and specific regulatory framework: SOLAS, MARPOL, ISM, MLC 2006, STCW, flag state requirements, port state control. The specific regulations and standards relevant to your business are also the terms your potential clients use when searching. A P&I correspondent who mentions MLC 2006 and port state control in their llms.txt will appear in responses to queries about those topics. One who does not, will not.

The Exact Format: What Your llms.txt Should Look Like

Here is a template adapted for a maritime services business. Replace the specifics with your own:

# CompanyName

> CompanyName is a [service type] based in [location], specialising in [specific service] for [specific client type]. [One additional sentence on geographic coverage or regulatory specialisation.]

## Services

### [Primary Service Name]

[Two to three sentences describing the service, including specific vessel types, regulatory frameworks, or client profiles it applies to.]

- [Link to service page]: [One sentence describing what this page covers]

### [Secondary Service Name]

[Description]

- [Link]: [Description]

## About

- [Link to about page]: Background on the company, founding, team, and certifications
- [Link to contact page]: Contact information and enquiry form

## Geographic Coverage

[Plain language description of your service territory]

## Regulatory Context

[The specific regulations, standards, and conventions your work relates to]

The file lives at yourdomain.com/llms.txt. It is a plain text file with a .txt extension. No special hosting configuration is required.

llms.txt vs llms-full.txt

The llms.txt standard also specifies an optional companion file: llms-full.txt. Where llms.txt is a structured index that links to your content, llms-full.txt is a consolidated full-text version of your most important content, formatted for AI consumption.

llms-full.txt is valuable for companies whose most important content is behind login walls, in PDFs, or otherwise not easily crawlable. By including the text of key documents in llms-full.txt, you make that content available to AI systems that might otherwise miss it.

For most maritime companies starting out, llms.txt is the priority. Build it first, verify it is accessible, and then consider whether llms-full.txt adds value based on your specific content situation.

How to Create and Deploy It in Under an Hour

Creating your llms.txt requires no technical skill beyond being able to create a text file and upload it to your web server. Here is the process:

Open any plain text editor. Write your company description, service sections, and supporting information following the format above. Save the file as llms.txt with UTF-8 encoding. Upload it to the root directory of your web server, the same level as your index.html or home page. Verify by visiting yourdomain.com/llms.txt in a browser. You should see the plain text content.

If your website is on WordPress, place the file in the root of your WordPress installation, not in the wp-content or themes directory. If you are on a hosted platform that does not allow root file access, check whether your platform supports custom file serving or static file hosting.

Once deployed, test it by using AI search tools to query for your services. Ask ChatGPT or Perplexity about the specific services you offer in your specific geography. Note whether your company appears. Run the same query three months after implementing llms.txt and compare the results. The difference will be measurable.

What Happens If You Do Not Have One

The absence of llms.txt does not mean you are completely invisible to AI systems. Models can still crawl your site and incorporate your content into their responses. But you are leaving the interpretation entirely up to the model, which means:

The model may read the wrong pages. The model may mischaracterise what you do. The model may assign lower confidence to its understanding of your business and therefore cite you less frequently. The model may accurately represent your 2019 blog post content rather than your current service offering.

llms.txt is not a magic switch that guarantees AI visibility. It is a signal of intentionality that AI systems are designed to prefer over sites that have not made the same effort. In a sector where almost no one has the file, having it is a material competitive advantage. That window will not stay open forever.

Check Your Current AI Visibility Score

The free Marine Brand Audit at Signe Agency checks whether your site has an llms.txt file, whether it is correctly structured, and how it contributes to your overall AI and GEO visibility score. The audit covers four other categories as well: technical SEO, content structure, conversion readiness, and traditional search signals. You receive a score out of 100 and three priority fixes to implement immediately.

Run it at signeagency.com/audit. It takes three minutes and the results are specific enough to act on today.