Go to documentation

Improve your Nuxt.js v2 Google Lighthouse score by delaying hydration.


NPM Downloads

Status: Stable v2 v0 ✅ , v3 main
Made possible by my Sponsor Program 💖
Follow me @harlan_zw 🐦


  • 🔥 Reduce your "Blocking Time" by as much as 100%, instantly increasing your Google Lighthouse score
  • 🚦 Filter it to run only on specific pages
  • 🍃 Minimal config
  • 🔁 (optional) Replay pre-hydration clicks

Why delay hydration?

Delaying hydration is a technique to hint to Google that our scripts are not required for our app to function.

By delaying hydration we improve the Google Lighthouse score by reducing your "Blocking Time" metric.

Previously you may have used vue-lazy-hydration which works well. However, it is just a more verbose way of providing hints to Google, like we're doing with this module.

What is a progressively enhanced app?

A progressively enhanced app is one that is designed to work without JavaScript, and then progressively enhanced with JavaScript.

Your scripts should not be required to use your website, this is what we're hinting to Google by delaying the hydration.

To do that you can ensure:

  • Full HTML served on the initial response
  • Scripts don't trigger a CLS
  • Avoid using scripts to set images, will affect the LCP

How this module works
A promise is injected into your app, the location is based on the mode. The promise is resolved as soon as either of these events has fired:
  • an interaction event (mouse move, scroll, click, etc)
  • an idle callback with a fixed timeout

The idle CPU time hints to Google that these scripts are not required for your app to run.

For example:

  • if a Google bot visits the page and has no interaction, out of the box the hydration won't occur until the browser idle callback + 6 seconds
  • if a user visits the page and moves their cursor or scrolls, the hydration will be triggered immediately. The chance of interacting with the non-hydration app will be minimised

Keep in mind, this is a hacky solution. Until Google can recognise progressive script enhancements, we'll need to rely on this approach.


If you're using Nuxt 2.x, please follow the docs on the v0 branch.

yarn add -D nuxt-delay-hydration# npm i -D nuxt-delay-hydration# pnpm add -D nuxt-delay-hydration

Requirement: Progressively enhanced SSR or SSG Nuxt app.

:warn: This module is experimental


// nuxt.config.tsexport default {  modules: [    'nuxt-delay-hydration',  ],  delayHydration: {    // enables nuxt-delay-hydration in dev mode for testing      debug: process.env.NODE_ENV === 'development'  }  }

Note: The module will not run in development unless you have enabled debug.

Choosing a mode

By default, no mode is selected, you will need to select how you would the module to work.

Type: init | mount| manual | false

Default: false

TypeDescriptionUse Case
false defaultDisable the moduleTesting
initDelays all scripts from loading.Zero or minimal plugins/modules.
mount recommendedDelays Nuxt while it's mounting. Plugins and some third-party scripts will work.Minimal non-critical plugins and third-party plugins.
manualDelay is provided by the DelayHydration component.All other apps

Regardless of the mode you choose, please read further optimisations.

Init Mode

This mode delays all scripts from loading until the hydration promise is resolved.

It does this by hooking into the HTML rendering, removing all script tags and adding them back after the hydration promise is resolved.

This will provide the biggest speed improvements however is the riskiest.

Pros: Provides the biggest blocking time reduction

Cons: Risky if you have critical third party scripts

Benchmark: ~90-100% reduction

export default {  delayHydration: {    mode: 'init'  }}

Mount Mode

This mode delays Nuxt while it's mounting. Plugins and some third-party scripts will work.

This delays your layout and page components.

Pros: Safer and still provides good improvements

Cons: May still break certain layouts if they are js dependent.

Benchmark: ~70% reduction

export default {  delayHydration: {    mode: 'mount'  }}

Manual Mode

Using the manual mode, you manually specify what part of your app you'd like to delay. Useful for when you need some part of the page to always hydrate immediately, such as a navigation drawer.

Pros: Safest way to optimise

Cons: Speed improvement based on usage

export default {  delayHydration: {    mode: 'manual'  }}

DelayHydration component

Once you have set the mode, you need to use the component.

<template><div>  <DelayHydration>    <div>        <LazyMyExpensiveComponent />    </div>  </DelayHydration></div></template>



Debug mode
It's recommended that you do thorough testing on your app with the module before deploying it into production.

To make sure the module is doing what you expect, there is a debug mode, which when enabled will log behaviour in the console.

It might be a good idea to always debug on your local environment, in that instance you could do:

export default {    delayHydration: {        debug: process.env.NODE_ENV === 'development',    },}

Visualising the hydration status
It can be unclear at times whether your app has been hydrated or not if it's quite static, this can make debugging hard.

To make things easier, there is a component HydrationStatus which will tell you what's going on.

<template><div>    <MyHeader />    <DelayHydration>        <div>            <!-- Show the hydration status, only for debugging -->            <HydrationStatus />            <main>                <nuxt />            </main>            <my-footer />        </div>    </DelayHydration></div></template>

Performance Auditing

Use my audit tool:

Replaying hydration click

What is this and how to enable
One of the issues with delaying hydration is that a user interaction event can occur before your scripts are loaded, leading to a user having to click on something multiple times for it to do what they expect. Think of a hamburger that is triggered using Javascript, if your app isn't hydrated then clicking it won't do anything.

The best fix for this is to write your HTML in a way that doesn't need Javascript to be interactive.

However, there are use cases where you need to use Javascript and responding to the first click is important. In those instances, you can enable the replay of the click.

export default {    delayHydration: {        replayClick: true    },}

This is an experimental configuration, you should test this option yourself before implementing it into your production app.

Further Optimisations

Load heavy components async
When you load in a heavy component synchronously, the javascript will be bundled in with the main application payload.

This will decrease all of your performance metrics. It's recommended you use async imports for these components.

Run nuxi analyze to find large components. When loading them, prefix them with Lazy.

Advanced Configuration

Configuration should be provided on the delayHydration key within your Nuxt config.

If you're finding the lab or field data is not performing, you may want to tinker with this advanced configuration.

Filtering routes

Using the include and exclude options, you can specify which routes you'd like to delay hydration on.

// nuxt.config.tsexport default {  delayHydration: {    include: [      '/blog/**',    ],    exclude: [      '/admin/**'    ],  },}

You can provide a glob pattern similar to route rules or a regex.

Event Hydration


  • Type: string[]
  • Default: [ 'mousemove', 'scroll', 'keydown', 'click', 'touchstart', 'wheel' ]

Controls which browser events should trigger the hydration to resume. By default, it is quite aggressive to avoid possible user experience issues.


  • Type: boolean
  • Default: false

If the trigger for hydration was a click, you can replay it. Replaying it will re-execute the event when it is presumed your app is hydrated.

For example, if a user clicks a hamburger icon and hydration is required to open the menu, it would replay the click once hydrated.

⚠️ This is experimental, use with caution.

Idle Hydration


  • Type: number (milliseconds)
  • Default: 7000

When waiting for an idle callback, it's possible to define a max amount of time to wait in milliseconds. This is useful when there are a lot of network requests happening.


  • Type: { mobile: number, desktop: number } (milliseconds)
  • Default: { mobile: 5000, desktop: 4000, }

How many to wait (in milliseconds) after the idle callback before we resume the hydration. This extra timeout is required to avoid the standard "blocking", we need to provide real idle time to lighthouse.

Mobile should always be higher than desktop as the CPU capacity will generally be a lot less than a desktop.

Note: The default will likely be customised in the future based on further benchmarking.



  • Type: boolean
  • Default: false

Log details in the console on when hydration is blocked and when and why it becomes unblocked.


Live examples




MIT License © 2022 - Present Harlan Wilton