TuxVoice 0.1.2

Fixes a clip cutting out every two seconds.

Starting TuxVoice while a copy was already running built a second audio engine. Each copy created its own tuxvoice_board node and each ran the PipeWire graph reconciler on its own two-second timer, so one kept adding the soundboard-to-headphones link and the other kept taking it away. On a live graph the link count oscillated on a two-second period; that is what the interruption was.

TuxVoice is now a single instance: launching it again raises the window that is already open.

Install or update:

flatpak install --from https://tuxvoice.vercel.app/tuxvoice.flatpakref

Already installed from that link? flatpak update picks this up.

TuxVoice 0.1.1

The first release Flathub can take: 0.1.0 was tagged before the application identifier changed, so its tree still says dev.tuxvoice.TuxVoice.

Changes since 0.1.0

• The application identifier is now io.github.Behruzbek34.TuxVoice, on a domain this project actually controls.
• Sharing a clip asks for a name and a picture first, rather than uploading under whatever the file happened to be called. Artwork is stored beside its sound and shown on the card.
• Uploads carry no credentials: they go through a relay that holds the key server-side, so nothing secret ships in the binary.
• Community cards show how long a clip runs and how long ago it was shared.
• A picture no longer decides how wide its card is, so a grid of cards is a grid of the same card.
• The Flatpak requires Flatpak 1.16, because --device=input silently widens to --device=all on anything older.

Install

git clone https://github.com/Behruzbek34/tuxvoice
cd tuxvoice
cargo build --release
./packaging/install-local.sh

Needs PipeWire, WirePlumber, GTK4, libadwaita and ffmpeg.

No distribution packages yet. An Arch package is written and builds cleanly, but the AUR has paused new account registration. A Flathub submission is open.

Known limits

• "Use as system microphone" cannot work in a Flatpak: the sandbox does not allow changing the machine's default input. The setting says so rather than failing.
• Global shortcuts read /dev/input, so your user must be in the input group.

TuxVoice 0.1.0

First release.

A voice changer and soundboard for Linux. It puts a virtual microphone into the PipeWire graph, so Discord, a game or a browser hears your processed voice and your clips without knowing TuxVoice exists.

Effects — twenty of them, chained in any order: pitch with formant correction, formant, vocoder, auto-tune, harmonizer, equalizer, filter, distortion, bitcrush, ring modulator, chorus, flanger, phaser, vibrato, tremolo, echo, reverb, compressor, noise gate, gain. Fourteen presets to start from, and your own chains saved under a name, an icon and a shortcut key.

Soundboard — your own clips, plus a catalogue searched from inside the app. Bind a clip to a key and fire it while a game has focus.

Routing — choose what the other side hears, monitor yourself on a device of your choosing, and have the graph repaired every two seconds so a Bluetooth profile switch does not drop you out of a call.

Install

Build it:

git clone https://github.com/Behruzbek34/tuxvoice
cd tuxvoice
cargo build --release
./target/release/tuxvoice

Needs PipeWire, WirePlumber, GTK4, libadwaita and ffmpeg.

./packaging/install-local.sh puts the build in your application menu.

No packages yet. An Arch package is written and builds cleanly, but the AUR has paused new account registration, so there is nothing to install from until that reopens. A Flatpak is being prepared.

Known limits

• "Use as system microphone" cannot work in a Flatpak: the sandbox does not allow changing the machine's default input.
• Global shortcuts read /dev/input, so your user must be in the input group.