EntityMap: A Sitemap For What Your Site Knows

What is it? EntityMap is a proposed open standard that lets publishers declare entities, relationships, and source evidence for AI systems. Here is why it matters for search, SEO, and publisher attribution.

Search engines have had sitemaps for years. A sitemap tells crawlers where the pages are. EntityMap wants to answer a different question: what does this site actually know?

EntityMap is a proposed open standard for publishing a structured, entity-first index of a website’s knowledge. Instead of leaving AI systems to scrape pages, slice paragraphs, and guess relationships, EntityMap gives publishers a file that declares entities, definitions, relationships, source URLs, and attribution in a machine-readable format.

The simple pitch is this:

sitemap.xml tells crawlers what pages exist. entitymap.json tells AI systems what a site knows.

Why Does This Matters Now?

AI search systems, AI Overviews, answer engines, and RAG pipelines often work at the page and passage level. They fetch HTML, break content into chunks, and retrieve the pieces that look useful.

That process creates three publisher problems:

• Disambiguation: The same concept can appear under several names and get treated as separate ideas.
• Attribution: Publisher identity can get stripped away when passages are stored, summarized, or reused.
• Relationships: Connections between people, products, concepts, methods, and evidence remain buried in prose.

EntityMap tries to solve those issues by letting publishers declare their own knowledge graph in a predictable format.

What An EntityMap Contains

An EntityMap uses two root-level files:

https://example.com/entitymap.json
https://example.com/entitymap.html

The JSON file is the machine-readable source. The HTML file is the human-readable and crawler-readable companion page.

A basic entity entry can include:

• An entity ID
• An entity type
• A name
• A short description
• Alternate names or surface forms
• Relationships to other entities
• Evidence chunks from source pages
• Canonical source URLs
• Publisher attribution
• Freshness data

In plain English, it says: this is the thing we cover, this is what we call it, this is what it means, this is how it connects to other things, and this is the source page that proves it.

SEOs Should Pay Attention

Entity-first SEO has been discussed for years, but implementation has usually depended on schema, internal linking, knowledge panels, Wikidata, topical maps, and content structure.

EntityMap adds a more direct layer. It gives site owners a dedicated file for declaring:

• What topics the site has authority on
• Which terms are proprietary or publisher-defined
• Which entities connect to which other entities
• Which pages contain the supporting evidence
• Which publisher should be attached to extracted chunks

That last point is the big one. In AI retrieval, a chunk can travel far from the original page. If publisher identity does not travel with it, attribution weakens. EntityMap puts publisher identity directly into each evidence chunk.

This Is Not Not Not A Replacement For Schema

EntityMap should not be viewed as a replacement for schema markup, XML sitemaps, RSS, llms.txt, or normal HTML. It sits beside them.

Format	Primary Purpose
sitemap.xml	Lists crawlable URLs
Schema markup	Adds structured data to individual pages
robots.txt	Gives crawler access instructions
llms.txt	Points AI systems toward selected content
entitymap.json	Declares site knowledge, entities, relationships, and evidence

Think of it as the site owner’s declared map of meaning.

Where This Gets Interesting

The useful part is not merely listing entities. The useful part is declaring relationships.

For example, a publisher can state that one entity:

• Is part of another entity
• Measures another concept
• Improves an outcome
• Targets an audience
• Is affiliated with an organization
• Is regulated by a standard or rule

That gives AI systems more than a pile of paragraphs. It gives them structured claims with supporting evidence.

For publishers, that is a way to reduce guesswork. For AI systems, it offers cleaner retrieval. For SEOs, it opens a new technical layer between content strategy and machine consumption.

Publisher Attribution Is The Core SEO Angle

The EntityMap specification requires the publisher field inside each chunk to match the root publisher name exactly. That is not a tiny formatting detail. It is the attribution mechanism.

AI systems often store extracted passages in vector databases. Once content is chunked, the original page context can be weakened or lost. EntityMap keeps the publisher name attached to each chunk so attribution has a better chance of surviving retrieval and summarization.

That matters for any site trying to be cited, named, or trusted inside AI-generated answers.

Who Should Experiment With It First?

EntityMap makes the most sense for sites with clear subject matter depth and defined knowledge assets.

Good early fits include:

• Publishers with large educational libraries
• SaaS companies with product, feature, and methodology pages
• Medical, legal, and financial sites with reviewed content
• Conference and event sites with speakers, sessions, topics, and sponsors
• Local directories with businesses, services, cities, and categories
• Research-heavy B2B sites with original frameworks and named concepts

For a thin affiliate site or a lightly maintained brochure site, EntityMap is probably overkill. For a serious publisher, it is worth testing.

What To Watch

EntityMap is still early. The standard can be published and validated, but adoption by major AI systems is the real question.

The practical test will be simple:

• Will AI crawlers read it?
• Will retrieval systems trust it?
• Will publisher attribution improve?
• Will tools generate and maintain it cleanly?
• Will validators reduce spam and low-quality self-declared claims?

Those questions decide whether EntityMap becomes a useful technical standard or another good idea waiting for adoption.

SEW Take

EntityMap is worth watching because it addresses one of the biggest problems in AI search: AI systems consume content, but they often do not preserve the publisher’s structure, intent, or attribution.

SEO has spent decades helping crawlers find pages. The next fight is helping AI systems understand meaning without flattening the source into anonymous text fragments.

EntityMap is one proposed answer. It gives publishers a way to say, in machine-readable form: here is what we know, here is how it connects, and here is the evidence.

That is a technical SEO idea with real teeth.

FAQ

What is EntityMap?

EntityMap is a proposed open standard for publishing a structured index of a website’s entities, relationships, evidence chunks, and publisher attribution for AI systems.

Is EntityMap the same as an XML sitemap?

No. An XML sitemap lists URLs. EntityMap describes what a site knows and where the evidence for that knowledge lives.

Does EntityMap replace schema markup?

No. Schema markup describes data on individual pages. EntityMap provides a site-level knowledge index that connects entities, evidence, and relationships.

Where does the EntityMap file live?

The proposed convention is to publish entitymap.json and entitymap.html at the root of the domain.

Why does EntityMap matter for AI search?

AI systems often retrieve content as isolated chunks. EntityMap keeps entities, evidence, relationships, and publisher identity connected in a structured file.

Should publishers add EntityMap now?

Technical publishers, SaaS companies, directories, research sites, and content-heavy brands should consider testing it. Broad adoption by AI systems will determine its long-term value.

EntityMap Wants To Give AI Search A Sitemap For Meaning

Meta title: EntityMap: A Sitemap For What Your Site Knows

Links

• GitHub Open Source Project
• https://entitymap.org/