DEV Community

Jean-Christophe Viau
Jean-Christophe Viau

Posted on

OpenFeeder: Make Your Website LLM-Native

#ai #opensource #webdev #llm

LLMs are browsing the web. Are your pages ready for them?

Right now, AI agents, RAG pipelines, and LLM-powered search tools are scraping your website. They're parsing raw HTML, guessing what's content vs navigation, stripping CSS, losing structured data — and getting a degraded, noisy version of what you actually built.

There's a better way. It's called OpenFeeder.

What is OpenFeeder?

OpenFeeder is an open standard that lets websites expose their content natively to LLMs — structured, clean, real-time, and server-side. No scraping. No guessing. Just your content, exactly as you want it served.

Two endpoints. That's it:

https://yoursite.com/.well-known/openfeeder.json  ← discovery (metadata, capabilities)
https://yoursite.com/openfeeder                   ← content feed (articles, pages, products)
Enter fullscreen mode Exit fullscreen mode

An LLM agent hits the discovery endpoint, learns what your site offers and how to query it, then pulls structured content from the feed. Clean JSON. Proper types. No HTML soup.

Why does this matter?

Think about robots.txt — a simple convention that let search crawlers know where to go. OpenFeeder is the same idea, but for LLMs. Instead of letting bots scrape randomly and misunderstand your content, you give them a first-class interface.

Benefits:

  • Accurate content — LLMs get the real text, not nav menus and cookie banners
  • Structured metadata — titles, dates, authors, categories, tags — all there
  • Real-time — always fresh, not cached by some third-party scraper
  • You control it — choose what to expose, at what depth

Drop-in adapters for your stack

OpenFeeder ships with ready-made adapters for the most common platforms and frameworks, all with 21+ tests each:

Platform Status
WordPress ✅
Drupal ✅
Joomla ✅
Next.js ✅
Vite ✅
Express ✅
FastAPI ✅
Astro ✅
WooCommerce ✅

Quick example (Express adapter) 

const { openfeederExpress } = require('@openfeeder/express');

app.use(openfeederExpress({
  site: {
    name: "My Dev Blog",
    description: "Articles about web development",
    url: "https://myblog.dev"
  },
  content: async () => {
    // return your posts/pages as structured objects
    return await db.getPosts();
  }
}));
Enter fullscreen mode Exit fullscreen mode

That's it. Your site now has /.well-known/openfeeder.json and /openfeeder — ready for any LLM agent that speaks the protocol.

The bigger picture

We're heading toward a web where AI agents are first-class visitors. They browse, they read, they summarize, they take action. If your site isn't LLM-native, you're invisible to this wave.

OpenFeeder is the standard that makes it happen. Open-source. Framework-agnostic. Dead simple to adopt.

👉 Star the repo and add OpenFeeder to your stack: github.com/jcviau81/openfeeder

PRs welcome. Adapters for more platforms coming soon.

Top comments (0)