A minimalistic, highly resource-efficient virtual microphone application for Linux. Mix arbitrary audio sources (like RTSP streams from IP cameras) into a single virtual microphone.

Designed for a Speech-Input Agnostic workflow. If you use persistent audio sources (like a room camera) for coding by voice, OpenMic provides a stable, always-on system microphone without the overhead of heavy streaming software.

• Linux with PipeWire or PulseAudio
• ffmpeg installed
• Python 3

1. Clone the repository:

   git clone https://github.com/shured/openmic.git
   cd openmic

2. Make the script executable:

   chmod +x openmic.py

3. Copy the example config to your active config directory:

   mkdir -p ~/.config/openmic
   cp config.json.example ~/.config/openmic/config.json

4. Adjust ~/.config/openmic/config.json with your real RTSP camera URL.
5. Auto-start on boot using systemd:

   # Copy service file
   mkdir -p ~/.config/systemd/user/
   cp openmic.service ~/.config/systemd/user/
   
   # Enable and start
   systemctl --user enable --now openmic.service

See ~/.config/openmic/config.json. You can adjust:

• virtual_mic_name: The name of the virtual microphone that will show up in your settings and chat applications.
• sources: A list of streams. Add as many as you need to mix them together.

{
    "virtual_mic_name": "OpenMic",
    "sources": [
        {
            "name": "ipcam",
            "url": "rtsp://admin:[email protected]:554/live/ch0",
            "options": ["-rtsp_transport", "tcp"]
        }
    ]
}

1. Creates a module-null-sink (mixer) via PulseAudio/PipeWire APIs.
2. Creates a module-virtual-source (the microphone device) acting as a monitor for that sink.
3. Spawns an ffmpeg instance for each configured stream to securely and efficiently output audio into the mixer sink.
4. If a stream drops or fails, the tool automatically tries to restart it.

MIT License