On-demand Rendering Adapters

Astro allows you to choose on-demand rendering for some, or all of your pages and endpoints. This is also known as server-side rendering (SSR): generating HTML pages on the server when requested and sending them to the client. An adapter is used to run your project on the server and handle these requests.

This on-demand rendering allows you to:

• Implement sessions for login state in your app.
• Render data from an API called dynamically with fetch().
• Deploy your site to a host using an adapter.

Consider enabling on-demand server rendering in your Astro project if you need the following:

• API endpoints: Create specific pages that function as API endpoints for tasks like database access, authentication, and authorization while keeping sensitive data hidden from the client.

• Protected pages: Restrict access to a page based on user privileges, by handling user access on the server.

• Frequently changing content: Generate individual pages without requiring a static rebuild of your site. This is useful when the content of a page updates frequently.

Official Adapters

Section titled Official Adapters

Astro maintains official adapters for Node.js, Vercel, Netlify, and Cloudflare.

Find even more community-maintained adapters (e.g. Deno, SST, AWS) in our integrations directory.

SSR 配接器

• 

  @astrojs/cloudflare

• 

  @astrojs/netlify

• 

  @astrojs/node

• 

  @astrojs/vercel

Enable on-demand server rendering

Section titled Enable on-demand server rendering

Both of Astro’s on-demand rendering output modes (server and hybrid) allow you to take advantage of static site performance by pre-rendering individual routes whenever possible, whether you have an entirely dynamic app or a mostly static site that needs on-demand rendering only for select routes.

To decide which one to use in your project, choose the output option that represents how most of your pages and routes will be rendered:

• output: 'server': On-demand rendered by default. Use this when most or all of your site or app should be server-rendered on request. Any individual page or endpoint can opt-in to pre-rendering.
• output: 'hybrid': Pre-rendered to HTML by default. Use this when most of your site should be static. Any individual page or endpoint can opt-out of pre-rendering.

Because the server will need to generate at least some pages on demand, both of these modes require you to add an adapter to carry out the server functions.

Add an Adapter

Section titled Add an Adapter

To deploy a project in server or hybrid mode, you need to add an adapter. This is because both of these modes require a server runtime: the environment that runs code on the server to generate pages when they are requested. Each adapter allows Astro to output a script that runs your project on a specific runtime, such as Vercel, Netlify or Cloudflare.

You can find both official and community adapters in our integrations directory. Choose the one that corresponds to your deployment environment.

astro add install

Section titled astro add install

You can add any of the official adapters maintained by Astro with the following astro add command. This will install the adapter and make the appropriate changes to your astro.config.mjs file in one step.

For example, to install the Vercel adapter, run:

npm

npx astro add vercel

pnpm

pnpm astro add vercel

Yarn

yarn astro add vercel

Manual Install

Section titled Manual Install

You can also add an adapter manually by installing the package and updating astro.config.mjs yourself.

For example, to install the Vercel adapter manually:

1. Install the adapter to your project dependencies using your preferred package manager:

   npm

   npm install @astrojs/vercel

   pnpm

   pnpm add @astrojs/vercel

   Yarn

   yarn add @astrojs/vercel

2. Add the adapter to your astro.config.mjs file’s import and default export, along with your desired output mode:

   astro.config.mjs

   import { defineConfig } from 'astro/config';
   import vercel from '@astrojs/vercel/serverless';
   
   export default defineConfig({
     output: 'server',
     adapter: vercel(),
   });

   Note that different adapters may also have different configuration settings. Read each adapter’s documentation, and apply any necessary config options to your chosen adapter in astro.config.mjs

Configure server or hybrid

Section titled Configure server or hybrid

To enable on-demand rendering, you must update your output configuration to one of the two server-rendered modes.

For example, to configure a highly dynamic app where every page is rendered on demand by default, add output: 'server' to your Astro config:

astro.config.mjs

import { defineConfig } from 'astro/config';
import node from "@astrojs/node";

export default defineConfig({
  output: 'server',
  adapter: node({
    mode: "standalone"
  })
});

Opting-in to pre-rendering in server mode

Section titled Opting-in to pre-rendering in server mode

For a mostly server-rendered app configured as output: server, add export const prerender = true to any page or route to pre-render a static page or endpoint:

src/pages/mypage.astro

---
export const prerender = true;
// ...
---
<html>
  <!-- Static, pre-rendered page here... -->
</html>

src/pages/mypage.mdx

---
layout: '../layouts/markdown.astro'
title: 'My page'
---
export const prerender = true;

# This is my static, pre-rendered page

src/pages/myendpoint.js

export const prerender = true;

export async function GET() {
  return new Response(
    JSON.stringify({
      message: `This is my static endpoint`,
    }),
  );
}

Opting out of pre-rendering in hybrid mode

Section titled Opting out of pre-rendering in hybrid mode

For a mostly static site configured as output: hybrid, add export const prerender = false to any files that should be server-rendered on demand:

src/pages/randomnumber.js

export const prerender = false;

export async function GET() {
  let number = Math.random();
  return new Response(
    JSON.stringify({
      number,
      message: `Here's a random number: ${number}`,
    }),
  );
}

On-demand rendering features

Section titled On-demand rendering features

HTML streaming

Section titled HTML streaming

With HTML streaming, a document is broken up into chunks, sent over the network in order, and rendered on the page in that order. In server or hybrid mode, Astro uses HTML streaming to send each component to the browser as it renders them. This makes sure the user sees your HTML as fast as possible, although network conditions can cause large documents to be downloaded slowly, and waiting for data fetches can block page rendering.

相關操作指南： Using streaming to improve page performance (EN)

Caution

Features that modify the Response headers are only available at the page level. (You can’t use them inside of components, including layout components.) By the time Astro runs your component code, it has already sent the Response headers and they cannot be modified.

Cookies

Section titled Cookies

In server and hybrid modes, a page or API endpoint can check, set, get, and delete cookies.

The example below updates the value of a cookie for a page view counter:

src/pages/index.astro

---
let counter = 0

if (Astro.cookies.has('counter')) {
  const cookie = Astro.cookies.get('counter')
  const value = cookie?.number()
  if (value !== undefined && !isNaN(value)) counter = value + 1
}

Astro.cookies.set('counter', String(counter))
---
<html>
  <h1>Counter = {counter}</h1>
</html>

See more details about Astro.cookies and the AstroCookie type in the API reference.

Response

Section titled Response

Astro.response is a standard ResponseInit object. It can be used to set the response status and headers.

The example below sets a response status and status text for a product listing page when the product does not exist:

src/pages/product/[id].astro

---
export const prerender = false; // Not needed in 'server' mode

import { getProduct } from '../api';

const product = await getProduct(Astro.params.id);

// No product found
if (!product) {
  Astro.response.status = 404;
  Astro.response.statusText = 'Not found';
}
---
<html>
  <!-- Page here... -->
</html>

Astro.response.headers

Section titled Astro.response.headers

You can set headers using the Astro.response.headers object:

src/pages/index.astro

---
Astro.response.headers.set('Cache-Control', 'public, max-age=3600');
---
<html>
  <!-- Page here... -->
</html>

Return a Response object

Section titled Return a Response object

You can also return a Response object directly from any page using on-demand rendering either manually or with Astro.redirect.

The example below looks up an ID in the database on a dynamic page and either it returns a 404 if the product does not exist, or it redirects the user to another page if the product is no longer available, or it displays the product:

src/pages/[id].astro

---
export const prerender = false; // Not needed in 'server' mode

import { getProduct } from '../api';

const product = await getProduct(Astro.params.id);

// No product found
if (!product) {
  return new Response(null, {
    status: 404,
    statusText: 'Not found'
  });
}

// The product is no longer available
if (!product.isAvailable) {
  return Astro.redirect("/products", 301);
}
---
<html>
  <!-- Page here... -->
</html>

Request

Section titled Request

Astro.request is a standard Request object. It can be used to get the url, headers, method, and even body of the request.

In both server and hybrid mode, you can access additional information from this object for pages that are not statically-generated.

Astro.request.headers

Section titled Astro.request.headers

The headers for the request are available on Astro.request.headers. This works like the browser’s Request.headers. It is a Headers object where you can retrieve headers such as the cookie.

src/pages/index.astro

---
const cookie = Astro.request.headers.get('cookie');
// ...
---
<html>
  <!-- Page here... -->
</html>

Astro.request.method

Section titled Astro.request.method

The HTTP method used in the request is available as Astro.request.method. This works like the browser’s Request.method. It returns the string representation of the HTTP method used in the request.

src/pages/index.astro

---
console.log(Astro.request.method) // GET (when navigated to in the browser)
---

See more details about Astro.request in the API reference.

Server Endpoints

Section titled Server Endpoints

A server endpoint, also known as an API route, is a special function exported from a .js or .ts file within the src/pages/ folder. A powerful feature of server-side rendering on demand, API routes are able to securely execute code on the server.

The function takes an endpoint context and returns a Response.

To learn more, see our Endpoints Guide.