Audionaut open-source multitrack audio editor launch

Audionaut 0.1.0 released: a free, cross‑platform multitrack audio editor with AI‑agent support

Audionaut is now available as an open‑source desktop application for precise audio cutting, multitrack playlists, and clean exports without the overhead of a full DAW. It runs natively on Windows, macOS, and Linux, is licensed under GPL‑3 (or later) with an optional commercial license, and ships with a rich set of features aimed at music, podcast, and multitrack workflows.

Core capabilities are ready out of the box

Multitrack editing – Create regions, drop clips onto per‑track playlists, and export the resulting mix.
Precise cutting & crossfades – The UI lets you split clips, move them, and automatically generate crossfades when clips abut.
AI‑agent editing via MCP – Audionaut implements the Model Context Protocol (MCP), enabling Claude and other agents to edit a project. Each agent edit appears as a single undo step, preserving full transparency.
Stem separation – Integrated demucs.cpp (a C++ port of Meta’s Demucs) downloads model weights on first use and can isolate drums, bass, vocals, and other stems.
Audio analysis – Optional Essentia integration provides BIC segmentation, onset detection, and beat tracking for advanced analysis pipelines.
Command‑line interface (audionaut-cli) – Headless scripts and CI pipelines can create, import, analyze, auto‑edit, and export .audium projects. Every command emits a single JSON envelope ({"ok": true, "result": …}) on stdout, making it straightforward to chain operations.

Building and installing the app

Platform	Build method
macOS	Open `Audionaut/Builds/MacOSX/Audionaut.xcodeproj` in Xcode and build.
Windows	Open `Audionaut/Builds/VisualStudio2026/Audionaut.sln` in Visual Studio 2026.
Linux	Install development packages (e.g., `libasound2-dev`, `libjack-jackd2-dev`, `libwebkit2gtk-4.1-dev`, etc.) and run `make CONFIG=Release -j8` inside `Audionaut/Builds/LinuxMakefile/`.

The repository must be cloned recursively to fetch submodules:

git clone --recursive https://github.com/kvoltmer/Audionaut.git
# or after cloning
git submodule update --init --recursive

Essentia (audio analysis) and demucs.cpp (stem separation) are optional submodules. Essentia requires Python ≤ 3.11, pkg-config, and CMake 3.x; the build script build_essentia.sh handles static linking and patches for macOS/Apple Silicon.

Using AI agents with Claude

Audionaut’s MCP support lets Claude edit a project directly. After installing Node 18+:

claude mcp add audionaut -- npx -y audionaut-mcp

Keep the project open in Audionaut; each edit arrives as a single undo step. macOS users should store projects in ~/Music because the sandboxed app can only access entitlement‑covered locations.

Command‑line workflow example

A typical AI‑driven pipeline looks like:

audionaut-cli create song.audium --channels 2
audionaut-cli import song.audium take1.wav take2.wav --position 4.5
audionaut-cli analyze song.audium --types sbic,beat_degara
audionaut-cli auto-edit song.audium --track 0 --measures 4
audionaut-cli export song.audium -o mix.wav --sample-rate 48000 --bit-depth 24

Each command returns a JSON result and an exit code (0 = success, 1 = operation failed, 2 = usage error, 3 = feature unavailable). The same verbs can be invoked directly on the GUI binary for headless operation while a GUI instance remains open.

Community feedback highlights

Usability suggestions – Users appreciate the core editing experience but request menu‑driven file import, keyboard shortcuts for splitting, automatic crossfades when clips overlap, envelope editing, and built‑in effects. (Comment by @atentaten)
Podcast use case – A listener wants to automatically remove music sections that disturb sleep, indicating potential for AI‑driven batch editing. (Comment by @thenoblesunfish)
Feature roadmap – Several commenters ask about label import/export, region labeling, and a dedicated feature list to aid LLM integration. (Comment by @bita_nidir)
Platform comparisons – Users compare Audionaut to DAWs and lightweight editors like Audiomass, and inquire about latency with many tracks. (Comments by @0gs, @Jeeetendra, @sureMan6)
Packaging requests – A request for a Flatpak distribution was noted. (Comment by @qpiox)
Technical curiosity – One user asked why JUCE was chosen over iPlug2, given JUCE’s commercial licensing constraints. (Comment by @r2ob)

These insights suggest immediate priorities: adding keyboard shortcuts, improving crossfade automation, exposing label APIs, and providing alternative packaging formats.

Licensing and contribution model

Audionaut is dual‑licensed under GPL‑3 (or later) and a commercial license, allowing both open‑source use and proprietary integration. Contributions are welcomed; a Contributor License Agreement (CLA) must be signed, and the project follows standard GitHub contribution guidelines.

Where to get it

Download page: https://audionaut.app/download
Source repository: https://github.com/kvoltmer/Audionaut
Documentation: User manual, CLI README, and build instructions are in the docs/ and Audionaut/ directories of the repo.

Bottom line: Audionaut delivers a lightweight, cross‑platform multitrack editor with modern C++ performance, AI‑agent integration, and a powerful CLI, filling a niche for open‑source audio production tools while inviting community‑driven enhancements.