wavesurfer.js
Wavesurfer.js is an interactive waveform rendering and audio playback library, perfect for web applications. It leverages modern web technologies to provide a robust and visually engaging audio experience.
Table of contents
- Getting started
- API reference
- Plugins
- Advanced / reactive API
- CSS styling
- Frequent questions
- Development
- Tests
- Feedback
Getting started
Install and import the package:
npm install --save wavesurfer.js
import WaveSurfer from 'wavesurfer.js'
Alternatively, insert a UMD script tag which exports the library as a global WaveSurfer variable:
<script src="https://unpkg.com/wavesurfer.js@7"></script>
Create a wavesurfer instance and pass various options:
const wavesurfer = WaveSurfer.create({
container: '#waveform',
waveColor: '#4F4A85',
progressColor: '#383351',
url: '/audio.mp3',
})
To import one of the plugins, e.g. the Regions plugin:
import Regions from 'wavesurfer.js/dist/plugins/regions.esm.js'
Or as a script tag that will export WaveSurfer.Regions:
<script src="https://unpkg.com/wavesurfer.js@7/dist/plugins/regions.min.js"></script>
TypeScript types are included in the package, so there's no need to install @types/wavesurfer.js.
See more examples.
Documentation
See the wavesurfer.js guide on our website for beginner-friendly, practical documentation covering common use cases, plugins, framework integration, and troubleshooting.
API reference
The full generated API reference lives at wavesurfer.xyz/docs/api:
Plugins
We maintain a number of official plugins that add various extra features:
- Regions – visual overlays and markers for regions of audio
- Timeline – displays notches and time labels below the waveform
- Minimap – a small waveform that serves as a scrollbar for the main waveform
- Envelope – a graphical interface to add fade-in and -out effects and control volume
- Record – records audio from the microphone and renders a waveform
- Spectrogram – visualization of an audio frequency spectrum (written by @akreal)
- Hover – shows a vertical line and timestmap on waveform hover
Advanced / reactive API
Beyond the imperative wavesurfer.on(...) events, v8 exposes a small reactive
surface for apps that want to read state instead of tracking it themselves:
const state = wavesurfer.getState()
state.isPlaying.subscribe((playing) => console.log('playing:', playing))
console.log(state.loadPhase.value) // 'idle' | 'fetching' | 'decoding' | 'ready' | 'error'
const { startTime, endTime } = wavesurfer.getRenderer().getVisibleRange().value
getState() returns read-only Signal objects
(.value to read, .subscribe(fn) to watch) for things like currentTime,
isPlaying, volume, muted, loadPhase and scrollPosition.
getRenderer().getVisibleRange() is a derived signal with the currently visible
{startTime, endTime} of the viewport, useful for plugins that need to sync to
what's on screen (e.g. a custom minimap or timeline).
If you're writing a plugin, WaveSurfer.definePlugin(name, (ctx, options) => api)
is an alternative to subclassing BasePlugin: ctx gives you { wavesurfer, scope, state, emit },
and anything you register on ctx.scope (listeners, timers, child scopes) is torn
down automatically on destroy() — no manual cleanup array required.
import WaveSurfer from 'wavesurfer.js'
const MyPlugin = WaveSurfer.definePlugin('MyPlugin', (ctx, options) => {
ctx.scope.listen(ctx.wavesurfer.getWrapper(), 'click', () => ctx.emit('my-event'))
return { doSomething: () => {} }
})
The Spectrogram plugin accepts
a rendering: 'windowed' option for very long audio files: it renders only the
visible time range and evicts off-screen segments, instead of computing the
whole file's frequency data up front.
SpectrogramPlugin.create({ rendering: 'windowed', /* ...other options */ })
This replaces the standalone WindowedSpectrogramPlugin (spectrogram-windowed.js),
which is now deprecated in favor of the merged option above — it's kept working
as a thin backward-compatible shim, but new code should use SpectrogramPlugin
directly.
CSS styling
wavesurfer.js v7 is rendered into a Shadow DOM tree. This isolates its CSS from the rest of the web page.
However, it's still possible to style various wavesurfer.js elements with CSS via the ::part() pseudo-selector.
For example:
#waveform ::part(cursor):before {
content: '🏄';
}
#waveform ::part(region) {
font-family: fantasy;
}
You can see which elements you can style in the DOM inspector – they will have a part attribute.
See this example to play around with styling.
Questions
Have a question about integrating wavesurfer.js on your website? Feel free to ask in our Discussions forum.
However, please keep in mind that this forum is dedicated to wavesurfer-specific questions. If you're new to JavaScript and need help with the general basics like importing NPM modules, please consider asking ChatGPT or StackOverflow first.
FAQ
I'm having CORS issues
Wavesurfer fetches audio from the URL you specify in order to decode it. Make sure this URL allows fetching data from your domain. In browser JavaScript, you can only fetch data eithetr from the same domain or another domain if and only if that domain enables CORS. So if your audio file is on an external domain, make sure that domain sends the right Access-Control-Allow-Origin headers. There's nothing you can do about it from the requesting side (i.e. your JS code).
Does wavesurfer support large files?
Since wavesurfer decodes audio entirely in the browser using Web Audio, large clips may fail to decode due to memory constraints. We recommend using pre-decoded peaks for large files (see this example). You can use a tool like audiowaveform to generate peaks.
What about streaming audio?
Streaming audio is supported only with pre-decoded peaks and duration.
There is a mismatch between my audio and the waveform. How do I fix it?
If you're using a VBR (variable bit rate) audio file, there might be a mismatch between the audio and the waveform. This can be fixed by converting your file to CBR (constant bit rate).
Alternatively, you can use the Web Audio shim which is more accurate.
How do I connect wavesurfer.js to Web Audio effects?
Generally, wavesurfer.js doesn't aim to be a wrapper for all things Web Audio. It's just a player with a waveform visualization. It does allow connecting itself to a Web Audio graph by exporting its audio element (see this example) but nothign more than that. Please don't expect wavesurfer to be able to cut, add effects, or process your audio in any way.
Why does my waveform look asymmetrical?
By default, wavesurfer.js renders stereo audio as a single waveform, with the left channel on the top half and the right channel on the bottom half. If your audio has hard-panned sounds, this will result in an asymmetrical waveform.
To render each channel separately, use the splitChannels option.
Development
To get started with development, follow these steps:
- Install dev dependencies:
yarn
- Start the TypeScript compiler in watch mode and launch an HTTP server:
yarn start
This command will open http://localhost:9090 in your browser with live reload, allowing you to see the changes as you develop.
Tests
The tests are written in the Cypress framework. They are a mix of e2e and visual regression tests.
To run the test suite locally, first build the project:
yarn build
Then launch the tests:
yarn cypress
Feedback
We appreciate your feedback and contributions!
If you encounter any issues or have suggestions for improvements, please don't hesitate to post in our forum.
We hope you enjoy using wavesurfer.js and look forward to hearing about your experiences with the library!
wavesurfer.js 是一个开源库，用于网页应用：它能渲染可交互的音频波形并播放音频，基于 Web Audio 等现代浏览器技术构建。