Audio feature extraction for JavaScript.

Getting Started

Meyda is a Javascript Library that can listen to audio and output a selection of statistics that describe it. Meyda has a variety of different applications, and environments it can be used in. Lets take a look at some.

What can you do with Meyda?

You can get audio analysis of

You can use these analyses in

What’s in these audio analyses?

Meyda can calculate a wide variety of standard audio features, including loudness, spectral characteristics like brightness (spectral centroid) and noisiness (flatness) and much more. For a full list and explanation of the audio features Meyda supports, please see our audio feature reference document.


To install Meyda in a modern frontend project, use npm:

npm install --save meyda

For web use cases, Meyda bundles a browserified copy of itself that exposes a Meyda object on the window object. This can be used in situations where npm is not in use, and you need to include a html link to your dependency, like the jQuery days. To add Meyda to your web page, you can load it via unpkg as follows:

<script type="text/javascript" src="<VERSION>/dist/web/meyda.min.js">

However, we recommend using npm to install Meyda where possible to avoid security issues, including Content Security Policy headers on your own site to allow Javascript sources from unpkg, and using subresource integrity to ensure the integrity of the installed script.

Typescript support

Type definitions are available in Meyda and are managed in the DefinitelyTyped repository. To get code completion in your IDE and type checking at compile time, install the @types/meyda package from npm alongside Meyda.

npm install --save @types/meyda

How can you do these things with Meyda?

Please have a look through the following tutorials for the one that closest matches your needs.

Need more help?

Can’t find a tutorial that matches up with what you’re trying to do with Meyda? Not finding the documentation to be clear enough? Please open a GitHub Issue, and we’ll do our best to help.

Fork me on GitHub