Mousai

Mousai is a music recognition application for the GNOME desktop, written entirely in Rust with GTK4 and libadwaita. It works like Shazam for your Linux desktop — hold it up to any audio source and it identifies the track playing within seconds. Mousai listens through your microphone or captures desktop audio directly via PipeWire, making it useful whether you caught a song on the radio or something is playing through your speakers that you can't quite name.

Features

• Microphone and desktop audio — Capture audio from your microphone or from system/desktop audio output via PipeWire, no external recording needed.
• AudD-powered recognition — Uses the AudD music recognition API, the same engine powering many mobile Shazam alternatives.
• Recognition history — Saves every successfully identified song so you can scroll back and find that track you recognised last week.
• In-app preview — Plays a short preview clip of the identified song directly inside the app so you can confirm the match.
• Open in streaming services — Jump straight to the identified track on Spotify, YouTube, or other services from the results view.
• Offline queue — Queue recognition requests while offline; Mousai will submit them automatically when connectivity is restored.
• Adaptive libadwaita UI — Follows GNOME design guidelines with full dark mode support and an adaptive layout that scales to different window sizes.
• Sandboxed and minimal — Ships as a Flatpak with network access scoped to only what recognition requires.

Installation

The easiest and recommended install method is Flatpak from Flathub:

flatpak install flathub io.github.seadve.Mousai

Mousai is also available via Snap:

snap install mousai

On Nix/NixOS:

nix-env -iA nixpkgs.mousai

For Arch Linux, install from the AUR using your preferred helper:

# Using paru
paru -S mousai

# Using yay
yay -S mousai

Mousai is not yet packaged in Debian/Ubuntu or Fedora repositories. The Flatpak from Flathub is the recommended route on those distributions.

Usage

Recognising a song

Launch Mousai and click the large Listen button. By default it listens through your microphone — hold your device near the audio source and wait a few seconds. The recognised song title, artist, album art, and links to streaming services will appear in the main view.

To recognise audio that is already playing through your system (desktop audio), open the app menu and switch the audio source to Desktop Audio before pressing Listen. This requires PipeWire.

Browsing history

All recognised songs are stored in the history list on the left. Click any entry to re-open the result and access its streaming links or preview.

AudD API key

Mousai includes a shared free-tier API key that provides a limited number of daily recognitions without any sign-up. For heavier use, register a free account at audd.io to receive your own API token with a higher quota. Enter the token in Preferences → AudD Token.

# Open Mousai preferences to set your AudD token
# Navigate to: hamburger menu → Preferences → AudD Token

Offline queue

If you tap Listen while offline, Mousai records the audio clip and queues it. When your internet connection returns, it automatically submits the queued clips and populates the history with the results.

AudD API and Privacy

Mousai sends a short audio fingerprint — not raw audio — to the AudD service for identification. No other data is transmitted. AudD's free tier allows a handful of recognitions per day per IP address; creating a free account at audd.io lifts this to a more comfortable daily limit. Paid plans are available for power users.

The app requests only the network Flatpak permission (scoped to the AudD API endpoint) and microphone or PipeWire access depending on the chosen source. No data is stored or shared beyond what is needed for recognition.

Mousai vs Other Song ID Tools

For a scriptable command-line song recognition tool, SongRec is a Rust-based alternative that uses the Shazam API instead of AudD.