Feature: FlyWeb convention for machine-readable content discovery

[sukrucan8]

What

FlyWeb is an open protocol that lets websites describe their structured data for AI agents. A site publishes /.well-known/flyweb.json:

{
  "flyweb": "1.0",
  "entity": "My App",
  "type": "docs",
  "resources": {
    "articles": {
      "path": "/.flyweb/articles",
      "format": "jsonl",
      "fields": ["title", "author", "date", "content"],
      "query": "?tag={tag}"
    }
  }
}

robots.txt tells machines where not to go. flyweb.json tells them what you have.

Why SvelteKit

SvelteKit already has patterns for serving structured files — +server.ts route files work perfectly for this:

// src/routes/.well-known/flyweb.json/+server.ts
import { json } from '@sveltejs/kit';

export function GET() {
  return json({
    flyweb: '1.0',
    entity: 'My App',
    type: 'blog',
    resources: {
      posts: {
        path: '/.flyweb/posts',
        format: 'jsonl',
        fields: ['title', 'author', 'date', 'content'],
        query: '?tag={tag}',
      },
    },
  });
}

A native integration could provide:

• Type-safe helper functions (import { flyweb } from '$app/flyweb')
• Auto-generation from your data structure
• Convention file support (like robots.txt in static/)

Context

AI agents are becoming primary consumers of web content. They scrape HTML, guess at structure, and hallucinate when wrong. FlyWeb gives them a standard discovery mechanism — like robots.txt but for structured data access.

Existing Implementation

• Spec: SPEC.md
• npm: flyweb (core + CLI validator)
• CLI: npx flyweb check https://example.com
• Next.js RFC: vercel/next.js#90075
• Homepage: flyweb.io