# Structured data vs llms.txt: a vocabulary engines read, a file agents might

> For the complete documentation index, see [llms.txt](https://attensira.com/llms.txt).
> Every page on this site is also served as Markdown: append `.md` to any URL, or send `Accept: text/markdown`.

> Structured data describes each page in a vocabulary search engines read; llms.txt lists your best pages for AI agents. Which one AI answers recommend, and why.

- Canonical page: https://attensira.com/compare/structured-data-vs-llms-txt
- Family: compare
- Last updated: 2026-10-04
- Author: Karl-Gustav Kallasmaa (https://attensira.com/about)
- Claims as data: https://attensira.com/claims.json?slug=structured-data-vs-llms-txt

**In short:** Structured data is schema.org markup, usually JSON-LD, inside each page, telling search engines what the page is; Google reads it for rich results. llms.txt is a proposed markdown file listing your best pages for AI agents. Google says neither is required for its AI features. If you do one, do structured data: it has documented consumers.

**Readings, 2 Oct 2026 to 4 Oct 2026**

| Measure | Value |
| --- | --- |
| stored AI answers that mention structured data or schema (24) (n=121 answers) | 19.8% |
| stored AI answers that mention llms.txt (6) (n=121 answers) | 5.0% |
| llms.txt mentions on prompts about developer docs | 3 of 6 |

_The newest stored ChatGPT and Perplexity answer to each of 61 tracked prompts, attensira.com workspace, read over the Attensira MCP on 4 Oct 2026 · 4 Oct 2026_

## Verdict

Consumed versus proposed. Structured data is a shared vocabulary that Google and other search engines document reading, embedded in every page. llms.txt is a single file proposed for AI agents, with no documented consumer among the answer engines. Neither is required for Google's AI features, by Google's own statement.

## The readings on this page

### The readings at the top

The newest stored ChatGPT and Perplexity answer to each of 61 tracked prompts, attensira.com workspace, read over the Attensira MCP on 4 Oct 2026 · 2 Oct 2026 to 4 Oct 2026

| Reading | Value | Sample |
| --- | --- | --- |
| stored AI answers that mention structured data or schema (24) | 19.8% | n=121 answers |
| stored AI answers that mention llms.txt (6) | 5.0% | n=121 answers |
| llms.txt mentions on prompts about developer docs | 3 of 6 |  |

### Which schema types do answers name?

Specific ones. Asked how to get docs cited by ChatGPT on 3 October 2026, Perplexity named FAQPage and Article schema and did not mention llms.txt, drawing on vendor and agency guides rather than OpenAI's documentation.

Stored Perplexity answer to “how to get my docs cited by ChatGPT”, asked 3 Oct 2026 (n=1 answer, attensira.com workspace, Attensira MCP).

> To get your docs cited by ChatGPT, make your content easy for AI to find and extract: publish direct, concise answers; use clear headings and structured data (schema); cite authoritative sources for factual claims; keep content up-to-date and browsable by search engines; and ensure pages are crawlable, fast, and clearly answer common questions in an FAQ-style format. Provide explicit, verifiable data and quotes where possible, and include FAQPage or Article schema.

Named, in order: no brand. Attensira was not named.

### What does ChatGPT read when asked how to get docs cited?

Its own vendor's pages. ChatGPT's 12 answers to "how to get my docs cited by ChatGPT" in the 30 days to 4 October 2026 carried 337 citations, and OpenAI domains supplied 337 of them. The schema advice came from Perplexity's third-party sources.

Citations in ChatGPT's answers to "how to get my docs cited by ChatGPT" · ChatGPT, 12 answers to the prompt "how to get my docs cited by ChatGPT" · n=337 citations · 5 Sep 2026 to 4 Oct 2026 · Deciding sources for one tracked prompt, read over the Attensira MCP, attensira.com workspace

| Name | Citations | Share |
| --- | --- | --- |
| help.openai.com | 283 | 84.0% |
| developers.openai.com | 42 | 12.5% |
| openai.com | 10 | 3.0% |
| help-lb.openai.com | 2 | 0.6% |

## How do Structured data and llms.txt compare?

Each cell cites a claim in the Sources section below.

| Criterion | Structured data | Llms.txt |
| --- | --- | --- |
| **What it is** |  |  |
| What it is | A standardised format for classifying page content | A proposal for a file that helps agents use a site |
| Who stands behind it | schema.org, naming Google, Microsoft, Yandex and Yahoo | A single author's proposal, first published September 2024 |
| Where it lives | Embedded in each page, usually as JSON-LD | A separate markdown file of links, at the root or a sub-path |
| **Evidence it is used** |  |  |
| Documented consumer | Yes · Google, for rich-result eligibility | No · None named; Google's Mueller called it purely speculative |
| Who publishes it | Any site that wants rich results, in a shared vocabulary | AI labs for their developer docs, per the proposal site |
| What correct implementation earns you | Eligibility for enhanced display, not a guarantee | No stated outcome; the proposal promises none |
| Required for Google's AI features | No special schema needed, per Google | No AI text files needed, per Google |
| **Our data, answers asked 2-4 Oct 2026** |  |  |
| Mentioned in AI answers (n=121 stored answers) | 24 answers | 6 answers, one advising against it |
| Perplexity on 'how to get my docs cited by ChatGPT' | Names FAQPage and Article schema | Not mentioned |

## What is Structured data?

A standardised format for describing what a page is about and classifying its content, expressed in the schema.org vocabulary and most often embedded in the page as JSON-LD, which Google recommends where a site can use it.

## What is llms.txt?

A proposal to standardise on a markdown file that gives AI agents a curated list of a site's most useful pages, published as a separate document at the site root or a sub-path rather than embedded in the pages themselves.

## When should you choose Structured data?

Invest in structured data when you want search engines to understand what each page is and be eligible for rich results. Generated from page data, it stays in step with the page, and it is what AI answers recommend most often in our readings.

## When should you choose llms.txt?

Add llms.txt when agents read your documentation directly and your platform can generate the file for you. It is a reading list for coding assistants and other agents, not a description of any page.

## When is neither the right choice?

Neither replaces a page that answers the question. Google states that no special schema.org structured data and no AI text files are needed for AI Overviews or AI Mode; what decides inclusion there is indexing, snippet eligibility and the page itself.

## Where does Attensira fit?

Attensira shows which of your pages each engine cites, per engine with n, before and after a markup or llms.txt change, so the change is judged on citations rather than on a validator's green tick.

## What sets Structured data and llms.txt apart?

- Only one of the two can drift out of step with a page unnoticed: markup is generated with the page, while a separate list of links goes stale when pages change.
- Their strongest cases are different audiences: structured data serves search engines reading any page, llms.txt serves agents reading developer documentation.
- This page carries the structured data it describes: FAQPage, BreadcrumbList and Claim JSON-LD generated from the visible text, and a markdown twin listed in our llms.txt.

## What is the difference between structured data and llms.txt?

Where each lives and who reads it. Structured data is markup inside each page that describes what the page is. Google defines it as a standardised format for providing information about a page and classifying its content, recommends JSON-LD where a site can use it, and reads it to decide whether a page is eligible for rich results. The vocabulary is schema.org, which describes itself as shared vocabularies the major search engines understand and names Google, Microsoft, Yandex and Yahoo.

llms.txt is a single markdown file at your site root, or at a sub-path for the pages beneath it, proposed by Jeremy Howard in September 2024. It lists the pages an AI agent would find most useful, with links and short descriptions, and points to clean markdown versions of them. It describes no page in detail. It is a reading list.

So one is a description attached to each page, in a vocabulary with documented consumers. The other is a curated index for a reader whose use of it the answer engines have not documented.

## What does each one do for AI answers?

For Google's AI features, neither is required. Google states that there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode, and that you don't need new machine-readable files, AI text files or markup. A page needs to be indexed and eligible to be shown with a snippet. Structured data still helps Google understand the page and can make it eligible for rich results in ordinary search, which is reason enough to have it.

For engines that fetch pages themselves, there is no published statement from OpenAI or Perplexity that either structured data or llms.txt changes which pages they cite. What they document is robots.txt for their crawlers. Google's John Mueller has gone further on llms.txt, calling it purely speculative for now and noting that the file has existed for years without the AI systems using it. Anything beyond that is inference from practice.

## What do AI answers themselves say about the two?

They mention structured data far more often, and they name specific types. We searched the newest stored answer to each of our 61 tracked prompts, one per engine, asked between 2 and 4 October 2026: 121 answers in all. Twenty-four mentioned structured data, schema markup or JSON-LD. Six mentioned llms.txt. Most of the schema mentions sat in lists of practical advice: add Organization, Product and FAQ markup, keep business details consistent, use clear headings. The llms.txt mentions clustered on prompts about developer documentation, MCP servers and tools that make site changes, which is where the file was designed to matter. One of the six, the ChatGPT answer above, told the reader not to over-invest in it.

Two cautions. These answers were assembled from third-party guides written by vendors and agencies, not from any engine's documentation, so they reflect what the industry writes rather than what the engine does when it chooses sources. And a mention is not a measurement of effect. Still, if the question is which of the two the engines' own answers steer people towards, the readings are clear.

## What should you spend your time on?

Pages first, structured data second, llms.txt third if it is free. A page that answers a buyer's question directly, near the top, with its numbers and their sources inside the passage, does more for AI answers than any markup. Then add structured data generated from the page's own data: Organization on the about page, Product or SoftwareApplication on product pages, Article on articles, FAQPage where the page really shows questions and answers, BreadcrumbList everywhere. Google requires every required property of a type before the markup makes a page eligible for an enhancement, so partial markup can be worth nothing.

If your platform generates llms.txt automatically, publish it. If it would mean hand-maintaining a list of URLs, skip it until someone documents a reason. Documentation platforms increasingly generate the file, and markdown versions of each page, without being asked, which removes the cost argument for docs sites but not the evidence question for everyone else.

## Are they solving the same problem?

No. Structured data answers "what is this page?" for a machine that is already reading it: this is a product, this is its price, these are the questions and answers on it, this is the organisation behind it. llms.txt answers "where should I start reading?" for a machine that has not yet chosen what to read.

That difference decides how each one fails. Markup generated from the same data as the page stays correct as the page changes. A list of pages in a separate file goes stale unless it is generated from the same source as the site, because nothing forces it to change when a page does.

## How many large sites publish llms.txt?

A minority. In our study of the top 10,000 domains of a dated Tranco list, run on 3 September 2026, we looked for an llms.txt file wherever the site's own robots.txt let our agent read it, and found one on 966 of the 7,781 domains checked. The method, the list and the per-crawler robots.txt results are on the study page. That is adoption without a documented reader, which is roughly where the proposal's own site and Google's comments leave it too.

## Why does structured data still matter when the reader is a model?

Because a model that fetches a page reads its HTML, and JSON-LD is part of it. A product's price in markup, an FAQ's questions and answers in markup, an article's author and date in markup: these are unambiguous statements a model can lift without parsing layout. Nobody has published proof that this raises citation rates, and we will not claim it does. But it costs little when generated from page data, it helps Google, and it removes one source of ambiguity for any reader.

## Do agents that read documentation behave differently?

Probably, and this is where llms.txt has its strongest case. Coding assistants and agents that work with a product's documentation often fetch documentation pages directly and benefit from a clean index and markdown versions, which is the use the proposal was written for. Its site says AI labs publish llms.txt files for their own developer docs and that Chrome's Lighthouse checks for one. Our readings agree: half of the llms.txt mentions came from prompts about developer docs. When we asked ChatGPT itself how to get docs cited, its 12 answers over 30 days drew all 337 of their citations from OpenAI's own domains, mostly the help centre: the engine answers that question from its vendor's documentation, not from the guides that sell either format. If your product has developer documentation, an llms.txt for the docs is a reasonable addition. If your site is marketing pages, the case is weaker.

### What does ChatGPT read when asked how to get docs cited?

Its own vendor's pages. ChatGPT's 12 answers to "how to get my docs cited by ChatGPT" in the 30 days to 4 October 2026 carried 337 citations, and OpenAI domains supplied 337 of them. The schema advice came from Perplexity's third-party sources.

**Citations in ChatGPT's answers to "how to get my docs cited by ChatGPT"** (ChatGPT, 12 answers to the prompt "how to get my docs cited by ChatGPT", 5 Sep 2026 to 4 Oct 2026)

| Name | citations |
| --- | --- |
| help.openai.com | 283 |
| developers.openai.com | 42 |
| openai.com | 10 |
| help-lb.openai.com | 2 |

n=337

_Deciding sources for one tracked prompt, read over the Attensira MCP, attensira.com workspace · 4 Oct 2026 · 337 citations in 12 answers_

## What does a good pairing look like?

This page, for example. Every compare page on our site carries FAQPage, BreadcrumbList and Claim JSON-LD generated from the same content as the visible text: the questions in the FAQ section are the questions in the FAQPage markup, and each sourced claim in the table is a Claim with its source URL and the date we read it. A markdown twin of the page sits under /md/ and is listed in our llms.txt. Nothing is described in markup that the page does not show.

That is the standard to aim for: markup and files that are by-products of the content, not separate artefacts someone has to remember to update.

## Which mistakes cost the most?

With structured data, three come up again and again. Markup that disagrees with the page, such as a price changed on the page but not in the template. Markup on the wrong page, such as Organization data repeated on every post with a different description each time, which gives a machine conflicting accounts of who you are. And partial markup that misses a required property, which Google treats as no markup for that enhancement. All three come from writing markup by hand, separately from the content.

With llms.txt, the costliest mistake is opportunity cost: a week spent perfecting the file while the pages bury their answers. The second is staleness, a file that lists removed pages or omits the ones you most want read. The third is believing it controls access. It does not; a page disallowed in robots.txt stays unfetchable however prominently llms.txt lists it.

## How can you tell whether either one made a difference?

By reading answers before and after, not by checking the markup. Pick the prompts the page should win, record how often each engine names you and cites the page over a few weeks, add or fix the markup, and keep reading. A green tick on a validator says the markup is well formed; it says nothing about citations. A difference that does not clear the noise is no difference, and on a small prompt set that usually means waiting longer than anyone wants to.

If someone tells you markup or a file raised their AI citations, ask how many answers the before and after figures rest on, which engines they read, and what else changed on the page in the same weeks.

## FAQ

### Does structured data help with AI answers?

It helps search engines understand a page, and Google uses it for rich results. Google states that no special schema.org structured data is needed to appear in AI Overviews or AI Mode. For engines that fetch pages themselves, there is no published statement that markup changes citation. Treat it as good hygiene, not a lever.

### Does llms.txt help with AI answers?

There is no documentation from OpenAI, Perplexity or Google saying their answer engines read it when choosing sources, and Google states its AI features need no AI text files. Its proposal site says AI labs publish llms.txt for their own developer docs and that Chrome's Lighthouse audits for one.

### What do AI answers themselves recommend?

Mostly structured data. Of the 121 newest stored ChatGPT and Perplexity answers to our 61 tracked prompts, asked 2 to 4 October 2026, 24 mentioned structured data, schema markup or JSON-LD and 6 mentioned llms.txt. One of those six told the reader not to over-invest in it.

### Which structured data types matter most?

The ones that describe what the page really is: Organization on your home or about page, Product or SoftwareApplication on product pages, Article on articles, FAQPage where the page shows questions and answers, BreadcrumbList everywhere. Google requires every required property of a type before the markup can make a page eligible for an enhancement.

### How common is llms.txt on large sites?

Still uncommon. When we checked the top 10,000 domains of a dated Tranco list in September 2026, an llms.txt file was found on 966 of the 7,781 domains whose robots.txt let us look. The full method is on our robots.txt study page.

### If I can only do one, which?

Structured data, generated from your page data. It has documented consumers in Google and other search engines and lives inside the pages it describes, so it cannot drift away from them. llms.txt is cheap to add later if your platform generates it.

## Sources

Each entry below is one claim on this page. `retrieved` is the date the exporter or a human last saw the source say this, and nothing more: it is not a publication date and not a promise of re-verification. `cite` is the anchor on the canonical page where this claim is rendered.

### sd-definition

Google defines structured data as a standardised format for providing information about a page and classifying the page content.

- source: https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data (developers.google.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-sd-definition
- side: structured-data

> a standardized format for providing information about a page and classifying the page content
### sd-jsonld

Google recommends JSON-LD for structured data where a site's setup allows it, as the easiest solution for website owners to implement and maintain.

- source: https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data (developers.google.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-sd-jsonld
- side: structured-data
### sd-eligibility-only

Google frames structured data as conferring eligibility rather than a guarantee, stating that all required properties must be included for an object to be eligible for enhanced display.

- source: https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data (developers.google.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-sd-eligibility-only
- side: structured-data
### schemaorg-backers

schema.org describes itself as providing shared vocabularies that webmasters can use to mark up their pages in ways the major search engines understand, and names Google, Microsoft, Yandex and Yahoo.

- source: https://schema.org/docs/gs.html (schema.org)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-schemaorg-backers
- side: structured-data
### llmstxt-purpose

llms.txt is described by its own site as a proposal to standardise on using an /llms.txt file to provide information to help agents use a website, authored by Jeremy Howard and first published in September 2024.

- source: https://llmstxt.org/ (llmstxt.org)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-llmstxt-purpose
- side: llms-txt

> A proposal to standardise on using an /llms.txt file to provide information to help agents use a website.
### llmstxt-markdown-format

The proposal specifies a markdown file of brief background, guidance and links to more detailed markdown pages, placed at the site root or at any path within it.

- source: https://llmstxt.org/ (llmstxt.org)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-llmstxt-markdown-format
- side: llms-txt
### llmstxt-publishers

The llms.txt site states that major AI labs, OpenAI, Anthropic and Gemini among them, publish llms.txt files for their own developer docs, and that Chrome's Lighthouse audits sites for one.

- source: https://llmstxt.org/ (llmstxt.org)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-llmstxt-publishers
- side: llms-txt
### mueller-no-consumers

Search Engine Journal reported that Google's John Mueller called llms.txt purely speculative for now, observing that the file has existed for years without the AI systems using it.

- source: https://www.searchenginejournal.com/google-says-llms-txt-is-purely-speculative-for-now/577576/ (www.searchenginejournal.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-mueller-no-consumers
- side: llms-txt
### google-no-special-schema

Google states that there is no special schema.org structured data you need to add to appear in AI Overviews or AI Mode, and no need for new machine-readable files, AI text files or markup.

- source: https://developers.google.com/search/docs/appearance/ai-features (developers.google.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-google-no-special-schema
- side: context
### dog-sd-mentions

Of the 121 newest stored ChatGPT and Perplexity answers to the 61 prompts tracked in the attensira.com workspace, asked 2 to 4 October 2026 and read over the Attensira MCP on 4 October 2026, 24 mentioned structured data, schema markup or JSON-LD.

- source: https://docs.attensira.com/mcp/tools (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-dog-sd-mentions
- side: structured-data
### dog-llms-mentions

Of the same 121 stored answers in the attensira.com workspace, read over the Attensira MCP on 4 October 2026, 6 mentioned llms.txt, one of them advising against over-investing in it.

- source: https://docs.attensira.com/mcp/tools (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-dog-llms-mentions
- side: llms-txt
### dog-pplx-docs

In the attensira.com workspace, read over the Attensira MCP on 4 October 2026, Perplexity's stored answer of 3 October 2026 to 'how to get my docs cited by ChatGPT' recommended FAQPage or Article schema and did not mention llms.txt.

- source: https://docs.attensira.com/mcp/tools (docs.attensira.com)
- retrieved: 2026-10-04
- cite: https://attensira.com/compare/structured-data-vs-llms-txt#claim-dog-pplx-docs
- side: context