HomeAverVOX OSS

The open-source front door

Private Linux voice—free and inspectable.

Install system-wide dictation, selected-text speech, spoken LLM conversations, Piper, Faster Whisper, the bridge CLI, and the warm daemon under the MIT License.

Open the GitHub repository Open on PyPI

MIT LicenseVersion 0.5.8PyPI aliases fixedNo telemetry

AverVOX OSSMIT licensed

Install

git clone https://github.com/avrvx/AverVOX-OSS.git
cd AverVOX-OSS
bash install.sh

# Python package alternatives
pipx install --system-site-packages avrvx
pipx install --system-site-packages avervox

pipx (or pip inside your own virtual environment) avoids the externally-managed-environment error a bare pip install hits on Ubuntu 24.04, Debian 12+, and Linux Mint 22. --system-site-packages lets AverVOX use the system PyGObject (python3-gi). You still need the system dependencies: GTK/PyGObject, PortAudio, xdotool, xclip, pulseaudio-utils (parec), and python3-dev (needed to compile evdev, a pynput dependency). Both avrvx and avervox commands are available starting with 0.5.5.

$0Price

MITLicense

0.5.8Current

Free for real work

Use OSS every day—and to test your setup.

OSS includes the desktop and developer workflows below, with no time limit. Pro adds what matters once voice is part of your day: natural Kokoro voices, a hands-free wake word, sessions that survive restarts, transcripts and dictation logs, and a LAN voice server.

Desktop voice

System-wide speech-to-text, selected-text TTS, global hotkeys, VAD, tray controls, conversation HUD, and barge-in.

LLM Converse

Streaming spoken conversations with any compatible endpoint, multiple profiles, sentence-level TTS, session headers, and failure detection.

Developer bridge

Synthesis, transcription, JSON capability discovery, streaming PCM, cancellation, a warm daemon, and integration installers.

Naming made explicit

Product name: AverVOX. Package and command: avrvx.

Version 0.5.5 also adds the fully spelled alias so users can install or launch the product using the name they heard.

Primary package: pip install avrvx
Alias package: pip install avervox → installs avrvx
Primary command: avrvx
Alias command: avervox

View avrvx on PyPI View alias package

Everyday commands

avrvx                    # launch tray app
avrvx --listen           # microphone → transcript
avrvx --speak "Hello"    # text → speakers
avrvx --capabilities     # JSON feature report
avrvx --daemon           # warm local bridge
avrvx --version

Running avrvx with no flags launches the tray application and registers the global hotkeys.

Interaction modes

Three hotkeys, plus a composable CLI.

Use only the parts you need. Dictation and selected-text speech do not require an LLM endpoint.

Dictate

Ctrl+Alt+Space
Speak into the focused application.

Read aloud

Ctrl+Alt+S
Hear the current text selection through Piper.

Converse

Ctrl+Alt+C
Hold a streaming spoken LLM conversation.

Automate

avrvx …
Use local speech in scripts and host applications.

OSS versus Pro

When voice becomes part of your day.

Pro keeps the same commands and integration contract, and adds hands-free activation, memory, history, nicer voices, and LAN sharing.

MIT licensed

AverVOX OSS

^$0

System-wide STT and TTS
Converse with streaming
Faster Whisper + Piper
Voice interruption and HUD
Bridge CLI + warm daemon
Published agent integrations

Install OSS

One-time license

AverVOX Pro

^$39 once

Everything in OSS
Kokoro voices + speed control
Wake-word activation
Persistent sessions + personas
LAN voice server + dashboard
Transcripts + dictation logs

Use the purchased version indefinitely. Includes all version 1 updates, up to but not including version 2.0.

Get Pro

OSS questions

Start in public. Keep control of the decision.

Why is the installer recommended over pip?

The repository installer also handles Linux system packages, the isolated environment, Piper voice model, desktop menu entry, autostart, and launcher. A bare pip install still requires the documented system dependencies, and on Ubuntu 24.04, Debian 12+, and Linux Mint 22 it fails outright with pip's externally-managed-environment error unless you use pipx install --system-site-packages avrvx or your own virtual environment created with --system-site-packages (so the system PyGObject is visible). The system dependencies are GTK/PyGObject, PortAudio, xdotool, xclip, pulseaudio-utils (parec), and python3-dev (needed to compile evdev, a pynput dependency); the installer handles all of them except pulseaudio-utils, which is preinstalled on most Mint and Ubuntu desktops.

Can OSS be used commercially?

The repository is published under the MIT License. Review the license file for the governing terms; the Pro license separately covers the commercial proprietary edition.

Where should OSS issues be reported?

Use the GitHub issue tracker for reproducible OSS problems, feature requests, and integration discussions. Licensed Pro customers may also use email support for product issues.

Is the OSS edition a trial that expires?

No. It is a separate open-source edition with no time limit. Pro is an optional one-time purchase that adds the documented premium features.

Can I contribute another integration?

Yes. The public integrations repository documents the bridge and existing host packages. A good integration should use capability discovery, preserve cancellation, prefer the warm daemon, and fall back safely.

Start with the workflow

Install the real product before making a purchase decision.

Use OSS to validate audio, hotkeys, text insertion, endpoint compatibility, and the voice workflow on your own Linux system.

Open the OSS repository Compare OSS and Pro