AI Domain Data (AIDD): A Standard for How AI Understands Your Site

Showcase
1

Hey everyone!

I want to share a new open standard I’ve created that I believe will matter a lot as AI systems become the primary way users discover and interact with websites.

It’s called AI Domain Data (AIDD). It provides every domain a clean, canonical, machine-readable way to describe who they are to AI systems, agents, and LLM-based search.

:link: Spec & tools:

https://github.com/ai-domain-data/spec
https://www.ai-domain-data.org

The Problem

Today, AI models and agents don’t have a reliable way to understand what a website is.
They guess from:

  • OpenGraph (often incomplete)

  • Schema.org (inconsistent)

  • About pages

  • Random metadata

  • Scraped HTML

This leads to:

  • Hallucinated descriptions of websites

  • Wrong attribution

  • Incorrect branding

  • Misclassification

  • Confusion between sites with similar names

As AI becomes a major traffic source (“AI SEO” is already becoming a very relevant topic), this is becoming a real problem for developers, marketers, founders, and products built on AI.

The Solution: AIDD

AIDD defines a simple, vendor-neutral JSON file that sits at:

/.well-known/domain-profile.json

It looks like this:

{

“spec”: “https://ai-domain-data.org/spec/v0.1”,

“name”: “Example Company”,

“description”: “Short, canonical description of your website.”,

“website”: “https://example.com”,

“logo”: “https://example.com/logo.svg”,

“contact”: “hello@example.com”,

“entity_type”: “business”, “blog”, “etc”

}

There’s also an optional DNS TXT mirror (_ai.yourdomain.com) for verification and fallback.

Think of it like a modern robots.txt, but for AI identity.

Why Developers Should Care

1. AI is becoming the primary discovery layer

Users increasingly get information through AI:

  • ChatGPT

  • Claude

  • Perplexity

  • LLM-driven search

  • Browser agents

  • On-device agents

You want those systems to describe your site correctly.

2. AIDD protects your brand from AI hallucinations

With AIDD, AI systems don’t have to guess your:

  • name

  • purpose

  • description

  • category

  • branding

This reduces misrepresentation and provides an authoritative source of truth.

3. Better visibility in AI agents and assistants

Agents can reliably index, categorize, and reference your domain.

This helps with “AI SEO”, the emerging equivalent of traditional SEO but for AI systems.

4. Zero vendor lock-in, tiny footprint, easy to host

  • Simple JSON

  • Lives in .well-known

  • No backend required

  • Works with static deployments

  • Vercel hosting is perfect for it

Deployment is literally one file.

5. AI platforms get huge benefits, meaning this is likely to scale

AI companies adopt AIDD because it gives them:

  • fewer hallucinations

  • lower compute cost

  • better grounding

  • stable provenance

  • cleaner knowledge graphs

And once they consume it, every website owner will want to publish it — just like sitemap.xml or OpenGraph.

Tools Available

  • Online generator

  • DNS helper

  • Validation tool

  • CLI (aidd)

  • Node/TypeScript resolver

  • GitHub Action

  • Cloudflare Worker & WordPress plugin coming soon

Everything is fully open-source.

I’d Love Feedback From the Vercel Community

  • Does this fit into your workflow?

  • Any fields you think should be added or removed?

  • Would you use a Vercel integration that auto-generates this?

  • Anything unclear or too opinionated?

I’m actively improving the spec and tooling and would love input from developers who build modern, AI-aware web experiences.

Thanks for reading, and if you try it, please share your thoughts!