audio.cpp is a pure C++ inference engine for audio models built on ggml, supporting TTS, STT, VAD, voice conversion, and more. Latest release adds Windows prebuilt binaries and GGUF loading; project activity includes recent commits and Docker/Build guidance.
What it is
audio.cpp is a high-performance C++ audio inference framework built on top of ggml, designed to make modern local audio models practical, portable, and fast.
How it works
It provides a portable native stack with CLI and server entry points (not Python-only paths), CUDA-optimized execution, reusable sessions, batch-style offline inference, and experimental JSON pipeline support. It includes built-in audio utilities (denoise, enhancement, resampling, STFT/ISTFT) and supports multiple backends (cpu, cuda, vulkan, metal).
Getting started
Linux Build
scripts/build_linux.sh --backend cuda --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend vulkan --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend cpu --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend cuda --native-cpu OFF --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend cuda --deployment-build --target audiocpp_cli --target audiocpp_server
Windows Build
powershell.exe -NoProfile -ExecutionPolicy Bypass -File .\scripts\build_windows.ps1
.\scripts\build_windows.ps1 -Preset windows-cuda-release -Target audiocpp_cli
.\scripts\build_windows.ps1 -Preset windows-cpu-release -Target audiocpp_cli
.\scripts\build_windows.ps1 -Target audiocpp_server -Jobs 16
### CLI usage
```bash
build/bin/audiocpp_cli \
--task <task> --model <path> [--family <family>] [--backend <backend>] [--mode <mode>] [options]
Common options include: --load-option, --session-option, --request-option, --config, --weight, --device, --threads.
Examples for TTS, ASR, and VC are provided in the README.
Docker
Docker CPU and CUDA images are available for both CLI and server use. See Docker.md for build commands and examples.