Astro middleware

Serve markdown twins from Astro middleware on on-demand rendered pages.

Astro middleware receives the request as context.request and renders the page through next().

import { defineMiddleware } from "astro:middleware";

import { twins } from "./lib/agentside";

export const onRequest = defineMiddleware((context, next) => twins.handle(context.request, () => next()));

import { agentside } from "@usenotra/agentside";
import { getEntry } from "astro:content";

export const twins = agentside({
  routes: {
    "/": "# Acme\n\nAcme sends invoices your customers actually pay.",
    "/blog/[slug]": async ({ params }) => (await getEntry("blog", params.slug))?.body ?? null,
  },
  skip: ["/_astro", "/_image"],
});

On-demand rendering

Middleware only sees real requests on pages that render on demand. Prerendered pages run middleware once at build time, without the visitor’s Accept or User-Agent headers, so negotiation cannot happen there. Use output: "server" with an adapter, or set export const prerender = false on the pages that should have twins.

.md URLs do not need their own pages. The middleware answers them before Astro looks for a route.