MorphMic

Real-time voice processor for Linux. Captures audio from a physical microphone, runs it through a DSP graph (gate, denoise, EQ, pitch / formant shift, ring modulator, band-pass, compressor, limiter, gain) and exposes the result as a virtual microphone MorphMic_Microphone that Discord, Zoom, OBS, the browser, etc. pick up via PipeWire/PulseAudio.

Components

The daemon and the clients talk over a Unix socket ($XDG_RUNTIME_DIR/morphmic/control.sock, mode 0600) using JSON-RPC 2.0.

License: GPL-3.0-or-later.

Installation

System dependencies (Ubuntu 24.04)

sudo apt install -y python3.12 python3-venv python3-pip \
    libportaudio2 portaudio19-dev librubberband-dev \
    pulseaudio-utils pipewire pipewire-pulse ffmpeg \
    autoconf automake libtool

# librnnoise is not in the Ubuntu 24.04 repos — build from source.
git clone --depth 1 https://github.com/xiph/rnnoise.git /tmp/rnnoise
(cd /tmp/rnnoise && ./autogen.sh && ./configure && make && sudo make install && sudo ldconfig)

Install MorphMic

Pick one. All three install the same morphmicd, morphmic-gui, morphmic-tui, morphmic console scripts into the venv's bin/.

# A) Editable install from a local clone (dev workflow)
git clone https://github.com/wachawo/morphmic.git
cd morphmic
python3.12 -m venv .venv
.venv/bin/pip install -e ".[gui,dev]"

# B) Install straight from GitHub (no clone)
python3.12 -m venv .venv
.venv/bin/pip install "morphmic[gui] @ git+https://github.com/wachawo/morphmic.git"

# C) Headless / CLI-only (skip the [gui] extra to avoid PySide6)
.venv/bin/pip install "morphmic @ git+https://github.com/wachawo/morphmic.git"

Extras: gui adds PySide6, dev adds pytest/ruff/black, all is gui+dev.

Docker Compose

docker compose up -d   # vc_daemon + vc_gui services

The host PipeWire socket is forwarded via /run/user/1000.

Quick start

After installing via pip (editable or from GitHub), activate the venv and use the short commands:

# 1. Start the daemon (creates MorphMic_Microphone, runs DSP with profile "clean").
morphmicd --profile clean

# 2a. Desktop GUI
morphmic-gui

# 2b. Console TUI (curses, works over SSH)
morphmic-tui

# 2c. CLI commands
morphmic status
morphmic profiles set deep
morphmic denoise on --level 0.7
morphmic echo on --delay-ms 1500

# Self-check
morphmic doctor

Equivalent dev forms (run straight from a checkout, no install needed):

python3 daemon/main.py --profile clean
python3 gui/main.py
python3 tui/main.py
python3 -m tui.cli status

In Discord/Zoom/OBS/the browser, pick MorphMic_Microphone as the input device.

Built-in profiles

Profiles are TOML files in ~/.config/morphmic/profiles/. Create custom ones from the GUI (New… → editor) or from the CLI (profiles export → edit → profiles save).

Audio engine

The daemon supports two backends:

• pipe (default when pw-cat or parec+pacat are installed) — capture/playback through subprocesses. Bypasses libpulse entirely; opens in ~50 ms instead of the 30 s timeout that PortAudio's pulse host API produces on PipeWire systems where pipewire-pulse is wedged.
• portaudio — sd.Stream(device=("pulse","pulse")). Lower latency, the original code path, recommended on classic PulseAudio.

Select the backend:

morphmicd --audio-engine pipe       # subprocess
morphmicd --audio-engine portaudio  # PortAudio
morphmicd --audio-tool pw-cat       # force PipeWire-native

Or in ~/.config/morphmic/config.toml:

audio_engine = "pipe"     # auto | portaudio | pipe
audio_tool = "auto"       # auto | pw-cat | parec

Testing

bash scripts/run_tests.sh   # ruff + black + pytest

192 unit + integration tests at v1.0.0.

System requirements

• Ubuntu 24.04 LTS or any distribution with PipeWire ≥ 1.0 + WirePlumber.
• Python ≥ 3.12.
• Any PipeWire-compatible microphone.

Project files

• ROADMAP.md — planned features for upcoming releases.
• CHANGELOG.md — release-by-release notes.
• LICENSE — full license text (GPL-3.0-or-later).