Next time you want to identify a song you hear in a TV show, movie, or other video give Mousai a go.

Mousai by SeaDve is a song recognition app for Linux desktop (and named after the ancient Greek Goddess of song and music). Built in GTK and leveraging the AudD song recognition API, Mousai is basically Shazam for Linux.

Open Mousai, hit the ‘listen’ button, play the song you want to identify (ideally in vague proximity to your laptop’s microphone) wait a few seconds, and bam: it tells you the song name and who performs it.

Mousai song recognition app listening on Linux desktop
Beep, boop

The app is ideal for identifying a song you heard in a TV show, advert, or movie, or when you’re out working from a coffee shop and the in-store radio plays a track you have to know more about.

Sometimes I’m watching a TV show on my laptop and I hear a song that I really like the sound of, but have no idea who sings it. My go-to method to find out who sings a song is pretty analog: memorise a few lines of lyrics then Google them with the word “lyrics” appended and hope there aren’t too many matching results.

But with Mousai I can select my desktop audio as the audio source, let the app listen, and discover what the song is called, and who performs it.

Mousai has a simple UI

Mousai uses the AudD API, which is rate limited. This means you can only use it to ‘listen’ to a handful of songs each day for free. Need more? You can register for your own API key and use it within the app. Be aware that AudD has its own privacy policy which you should check out prior to using this app, or signing up for API access.

A couple of additional ‘conveniences’ come baked into Mousai too, such as the ability to listen to excerpts of previously identified songs, and access a web page populated with links to play a song in full on popular music streaming services, including Spotify.

Quick link to stream the matched song

Mousai is entirely on-demand; it does not stay ‘listening’ in the background like similar apps for Windows and macOS do, which is a nice bonus.

So how well does it work?

I grabbed the latest version of the app from Flathub and tested it on my Fedora Rawhide (using GNOME 40.3) desktop. I got a 100% match rate when testing songs from established artists. From Spice Girls to New Found Glory it got each and every track I tried exactly right, first time, within seconds.

But what about when I tried less popular stuff?

Mousai is not perfect

Here results were mixed.

Mousai failed to identify tracks by some of the less mainstream pop-punk bands I listen to, but correctly matched others. Recency appeared to be the most indicative route to failure; the newer a song is the less likely it is to be identified by the AudD API.

If you plan to ‘test’ Mousai by throwing the most obscure b-side or lost recording you can find at it do be prepared to see the “Sorry. The song was not recognized” dialog box — and keep in mind that failed matches do use up your daily allowance of free matches.

Install Mousai on Linux

Mousai is free, open source software available from Flathub:

Get Mousai on Flathub

Source code and build instructions are up on GitHub should you fancy the challenge of running it with using the official Flatpak build.

thanks @KiveyGaming