open protocol v1.0

Make the internet
machine-readable.

An open protocol that lets any website describe its content for AI. One JSON file. Every machine understands it. Attribution enforced.

universal discovery structured data enforced attribution

Get Started Free Star on GitHub

/.well-known/flyweb.json

your-website.com

{
"flyweb": "1.0",
"entity": "Your Website",
"type": "blog",
"resources": {
"posts": {
"path": "/.flyweb/posts",
"format": "jsonl",
"fields": ["title", "author", "date", "content"]
}
}
}

1 file
your whole site
any AI
universal compat
5 min
to set up

The problem

The web wasn't built for machines.

And it's breaking everything AI tries to do.

</>

HTML Soup

Every website is a unique mess of divs, classes, and JavaScript. AI agents scrape raw HTML, guess what's content, and hallucinate when they get it wrong.

No Discovery

There's no standard way for AI to ask: "What content do you have?" Every API is bespoke. Every integration is custom. No universal protocol.

Zero Attribution

AI consumes your content, generates answers, and gives zero credit to the original creator. No link back. No control.

The result? AI that hallucinates, publishers that get scraped, and a web that no machine can reliably read.

The solution

One file. Everything changes.

Add /.well-known/flyweb.json to your website. It tells every AI what you have, how to read it, and how to credit you.

/.well-known/flyweb.json

{
  "flyweb": "1.0",
  "entity": "TechCrunch",
  "type": "news",
  "attribution": {
    "required": true,
    "license": "CC-BY-4.0",
    "must_link": true
  },
  "resources": {
    "articles": {
      "path": "/.flyweb/articles",
      "format": "jsonl",
      "fields": ["title", "author", "date", "content"],
      "access": "free",
      "query": "?tag={tag}&limit={n}"
    }
  }
}

Use my content. But say my name.

Attribution is enforced at the protocol level. Non-negotiable.

How it works

Three layers. One protocol.

Discovery

Your website publishes /.well-known/flyweb.json. AI agents check this file first — like robots.txt but the opposite. Instead of "stay away," it says "here's what I have."

Structure

Content is served as clean, structured data — JSON or JSONL. No more parsing HTML. No more guessing what's a headline vs. a sidebar ad. Just data.

Query

Standard URL queries any AI agent can construct. No SDK. No API key. Just GET /.flyweb/posts?tag=ai. Simple as a link.

Comparison

Before & After

Same content. Completely different experience for machines.

Without FlyWeb

Scraping HTML...

<div class="post-container mx-4">
  <div class="flex items-center gap-2">
    <img src="/avatars/sarah.jpg" />
    <span class="text-sm">Sarah Chen</span>
  </div>
  <h2 class="font-bold mt-4">
    AI Agents Need Structure
  </h2>
  <div class="prose mt-2">
    <p>The web was built for...</p>
    <div class="ad-banner">...</div>
  </div>
</div>
// Which part is the article?
// Where does content end?
// Is "Sarah Chen" the author?

With FlyWeb

GET /.flyweb/articles/ai-agents-need-structure

{
  "title": "AI Agents Need Structure",
  "author": "Sarah Chen",
  "date": "2026-02-15",
  "tags": ["ai", "web"],
  "content": "The web was built for...",
  "url": "https://example.com/posts/42"
}
// Clean. Structured. Zero guessing.