audio.cpp

audio.cpp

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.

Tired of juggling a dozen Conda environments, hundreds of Python packages, and dependency conflicts just to try a few audio models? audio.cpp gives those paths a shared native runtime instead. Runs on Windows, Linux, and macOS, with support for NVIDIA, AMD, Apple Silicon, and CPU-only machines.

Lost in a sea of audio model names, papers, and codebases? Explore 100+ audio models in the Audio Model Architecture Atlas to see how they're built, what they share, and where they differ, whether you're learning or building your own.

Supported backends:

[!TIP]

v0.9.4 highlights

Experimental HIP kernels: Select --backend hip --backend-impl gfx1151_v1 to try alternative AMD kernels. On the tested Strix Halo GPU, Q8 generation time fell by 76.7% for Irodori-TTS Small MF, 30.9% for Qwen3-TTS 1.7B VoiceDesign, and 14.0% for BreezeTTS2 versus default HIP. Despite its name, gfx1151_v1 can also run on RDNA1/2, as well as RDNA3/4; available fast paths vary by architecture. Gains vary by model and hardware; the implementation is opt-in. See the HIP guide. Contributions are welcome: test other AMD GPUs, report model performance and correctness, or help improve the kernels. Thanks to AMD DevLab for providing hardware access!

Metal optimizations: New SSM scan, convolution, and dataflow optimizations reduce RE-USE processing time by 56.5% on M4 Pro. Parakeet’s channel-contiguous subsampling and matrix projections reduce end-to-end processing time by 12.1–13.9%, with unchanged transcripts.

Browser inference with WASM + WebGPU: Run supported models locally in the browser. On RTX 5090 in Chrome, measured throughput reaches 23.3× real time for Supertonic 3, 13.9× for Kokoro, 94.3× for SenseVoice ASR, and 25.0× for warmed LFM3-Audio ASR. See the tested models, settings, and performance table.

Experimental streaming avatar voice chat: Talk to LFM3-Audio through the WebUI while SoulX-FlashHead animates your chosen portrait. Responses stream as text, speech, and synchronized head animation, with voice selection, adjustable speaking speed, and a customizable personality.

audio.cpp is built for real end-to-end execution rather than one-off model demos. Its optimized, reusable building blocks support speech generation and recognition, separation, alignment, and multi-model workflows. Shared components make new models easier to integrate and let improvements benefit multiple model families while keeping inference efficient, maintainable, and portable.

Features

⭐ 143 audio model families and 250+ variants. Run TTS, ASR, music generation, voice cloning, conversion, separation, and more locally through a shared CLI and server API.

⭐ Built-in WebUI. Generate and clone voices in Studio, compare models in Arena, manage models and voice libraries, and assemble narration with reusable speakers and passage editing in the TTS app.

⭐ Performance. Reusable sessions, cached graphs, and batched offline inference. Selected TTS paths run 1.8–8× faster than Python; tested Q8 packages deliver up to 1.53× speedup and 37% lower peak VRAM versus 16-bit weights (report). Supertonic 3 generates 10 hours of audio in 3 minutes on RTX 5090 (demo), while Nemotron 3.5 ASR matches WER at roughly ¼ the wall time in the TranscrIA benchmark. audio.cpp is among the fastest inference engines for streaming TTS latency on the Open TTS Leaderboard.

⭐ Portability. A native ggml runtime supporting CUDA, HIP/ROCm, Vulkan, Metal, WebGPU, and CPU through common CLI and server interfaces.

⭐ Hardware-specific optimizations. ARM NEON acceleration, Rockchip RKNN component execution, experimental HIP kernels, and CUDA, Metal, and Vulkan optimizations. Model-specific perf_mode presets expose optional fast paths for supported hardware.

⭐ Parity. Validation tooling for comparing model outputs against Python reference implementations.

⭐ Integrations and deployment. Powers speech, music, and transcription in Unsloth Studio. Fun-ASR-Nano and SenseVoice can also be deployed through the FunASR platform.

Contributing

audio.cpp would not be moving this quickly without generous contributors bringing in real fixes, new capabilities, and careful polish. See CONTRIBUTING.md for how to contribute and for a shout-out to the people already helping shape the project. Before starting a new model port, check the supported model table because several families are already implemented or under testing, and read the New Model PRs section in CONTRIBUTING.md. New ports should start under the community model source tree, where review is lighter than core models but still needs reproducible validation.

News

[!IMPORTANT] 2026-11-04 — Release v0.9.4: WebGPU support brings audio inference to the browser with WASM, and the WebUI gains streaming voice chat with animated SoulX avatars, cross-language dubbing, and voice-library export, import, and sorting. New models include Nemotron 3.5 Transcribe with CPU offload below 7 GB VRAM, LFM3-Audio, LFM D1-Omni, PocketTTS Small, SoulX-FlashHead, Real-time RE-USE, Magenta RealTime 2, Music Flamingo, Haas, and Underdog Bark 3, plus Irodori-TTS v4 Large. Experimental AMD kernel selection, new performance controls, and backend fixes improve speed, memory use, and compatibility. This release brings together community model ports, backend optimizations, and fixes shaped by real deployment reports.

2026-10-26 — Release v0.9.3: New speech models including Drama3, Ming 3.0 TTS/Audio Tiny, XTTS v2, Plapre 3, Gradium Phonon, and Earshot VAD, plus experimental LatentSync lip sync. Highlights include LoRA across 11 models, a TTS workspace, voice-cloning and dubbing recipes, perf_mode presets, lower memory use, and expanded HIP, ARM NEON, and Rockchip NPU support. Thanks to our contributors!

2026-10-15 — Release v0.9.2: New models include Voxtral TTS, TF-GridNet, MossFormer2, Hviske v6, FireRed VAD, AST, CED, and wake-word detection models. This release expands TTS streaming, adds multi-turn LFM2.5-Audio conversations, and improves model performance and compatibility. The WebUI gains a shared Voice Library and layout improvements. Downloadable C API libraries, Windows Vulkan portable builds, and packaged server configuration examples make integration easier. Thanks to everyone contributing code, testing, and feedback!

2026-09-30 - Release v0.9.0: 100+ audio models, an enhancement and denoising WebUI tab, speaker-tagged ASR, and performance and memory improvements. Explore the Architecture Atlas. Thanks to our community!

2026-06-25 to 2026-09-23 (releases 0.1 to 0.8): audio.cpp grew to 80+ model families and 120+ variants across TTS, ASR, music generation, voice conversion, separation, and speech analysis. Major additions included the native WebUI and Arena, GGUF packaging, LoRA adapters, streaming, and CUDA/Metal performance improvements, alongside Vulkan and experimental HIP support. The C ABI, Colab integration, MP3/HTTPS frontends, and models such as YuE2 and SheetSage2 expanded integration and audio-to-score workflows.

Supported Models

Task tags: TTS text to speech, Clone voice cloning, VC voice conversion, S2S speech-to-speech, ASR speech recognition, Classify audio classification, Wake wake-word detection, Align forced alignment, VAD voice activity detection, Diar speaker diarization, Codec audio codec, Sep source separation, MIDI audio-to-symbolic MIDI/events, Music music/song generation, SFX sound effects, Video video generation, Edit audio/music editing, Design voice design, Dialogue multi-speaker dialogue TTS, Ctrl TTS/clone voice control such as emotion, style, instruction, caption, or non-verbal tag control.

Runtime tags summarize the supported loading paths. GGUF package precision varies by model and release; check the audio.cpp GGUF repo or docs/gguf.md for the exact package list. Bundled means the tiny runtime asset ships under assets/framework/models and needs no separate model download. Stream means the family exposes a streaming server/session path. LoRA means adapter loading is supported in audio.cpp; see the LoRA guide for supported formats and usage.

Model weights keep the license of their original release, which is separate from audio.cpp's own license. docs/model_licenses.md lists that license per family, and whether it allows commercial use.

Speech Generation And Conversation

Family	Task	Lang	Variants	Runtime
breeze_tts	TTS, Clone, Design, Ctrl	zh, en	BreezeTTS 2 instruction-conditioned TTS and prompt-audio voice cloning	GGUF BF16/Q8/Q4_0, Stream
chatterbox	TTS, Clone, VC	ar, da, de, el, en, es, fi, fr, hi, it, ko, ms, nl, no, pl, pt, sv, sw, tr	Chatterbox with 0.5B backbone	GGUF 16/Q8
confucius4_tts	TTS, Clone	zh, en, ja, ko, de, fr, es, id, it, th, pt, ru, ms, vi	Confucius4-TTS multilingual voice cloning	GGUF F32, Stream
cosyvoice3	TTS, Clone	zh, en, ja, ko, de, es, fr, it, ru, yue	Fun-CosyVoice3 zero-shot, cross-lingual, and instruction-conditioned TTS	GGUF F32/Q8, LoRA
dots_tts	TTS, Clone, Edit, Ctrl	multilingual	DotTTS SOARDotTTS MeanFlowDotTTS Edit	GGUF 16/Q8, Stream
dramabox	TTS, Clone	en	DramaBox expressive TTS and voice cloning	GGUF Q8
fish_audio	TTS, Clone, Ctrl	auto, en, zh	Fish Audio S2 Pro	GGUF 16/Q8, Stream, LoRA
firered_audio	ASR, TTS, Clone, Design, Ctrl	zh, en	FireRedAudio multimodal speech/audio model with ASR, understanding, cloning, design, and edit paths	GGUF original/Q8
fireredtts3	TTS, Clone, Design, Ctrl	24 langs + 21 zh dialects	FireRedTTS3 BaseFireRedTTS3 Instruct/Voicedesign	GGUF original/Q8
gradium_phonon	TTS	en	Gradium Phonon 31m English	GGUF Q8, Safetensors, Stream
higgs_audio_tts	TTS, Clone, Ctrl	auto	Higgs Audio v3 TTS 4B	GGUF 16/Q8, Stream
index_tts2	TTS, Clone, Ctrl	zh, en, ja, es, ar	IndexTTS-2IndexTTS-2.5	GGUF 16/Q8
kokoro_tts	TTS	en-us, en-gb, es, fr, hi, it, ja, pt-br, zh	Kokoro 82M, 54 preset voices	Safetensors, local GGUF BF16/Q8, Stream
kugelaudio	TTS	European multilingual	KugelAudio-0-Open, four preset voices	GGUF BF16 / Q8_0 / Q4_K, Stream
irodori_tts	TTS, Clone, Design, Ctrl	ja	Irodori-TTS-v4.1-Small (RF and MF)Irodori-TTS-v4.1-AnimeIrodori-TTS-500M-v3Irodori-TTS-600M-v3-VoiceDesign	GGUF 16/Q8
magpie_tts	TTS	ar-AE, ar-MSA, ar-SA, de, en, es, fr, hi, it, ko, pt-BR, vi, zh	NVIDIA MagpieTTS Multilingual 357M (v2607) with baked speaker prompts and NanoCodec decode	GGUF original/Q8
maya1	TTS, Design, Ctrl	en	Maya1 expressive TTS with natural-language voice design, inline emotion tags, and SNAC decoding	GGUF original/Q8
ming3_tts	TTS, Clone, Design, Ctrl	zh, en, ja, ko, de, fr, es	Ming 3.0 TTS 0.5B	GGUF BF16/Q8, Stream
miotts	TTS, Clone	en, ja	MioTTS-1.7B	GGUF 16/Q8
moss_tts_local	TTS, Clone, Ctrl	auto, optional language hint	MOSS-TTS-Local-Transformer-v1.5	GGUF 16/Q8
moss_tts_nano	TTS, Clone	auto	MOSS-TTS-Nano-100M	GGUF 16/Q8, Stream
neutts	TTS, Ctrl	en	NeuTTS 2E with built-in speaker prompts and emotion control	GGUF original precision, Stream
omnivoice	TTS, Clone, Design, Ctrl	646+ langs	OmniVoice, Qwen3-0.6B basedVoiceTut-TTS (Egyptian Arabic fine-tune)	GGUF 16/Q8, Stream
personaplex	Dialogue, S2S	en	PersonaPlex 7B v1 speech-to-speech conversational model with packaged voice/persona prompts	GGUF Q4/Q8, Stream
plapre	TTS, Clone, Ctrl, Edit	da	Plapre 3, five preset voices	GGUF F32/BF16, Stream
pocket_tts	TTS, Clone	en, fr, de, it, pt, es	PocketTTS-100M English/French/German/Italian/Portuguese/Spanish	GGUF 16/Q8, Stream
pocket_tts_small	TTS, Clone	en	PocketTTS Small 18M/7M/2M, base and ReFL	GGUF F32, Stream
qwen3_tts	TTS, Clone, Design, Ctrl	zh, en, fr, de, it, ja, ko, pt, ru, es	Qwen3-TTS-12Hz-0.6B-BaseQwen3-TTS-12Hz-1.7B-BaseQwen3-TTS-12Hz-1.7B-CustomVoiceQwen3-TTS-12Hz-1.7B-VoiceDesign	GGUF 16/Q8, Stream, LoRA
supertonic	TTS	en, ko, ja, ar, bg, cs, da, de, el, es, et, fi, fr, hi, hr, hu, id, it, lt, lv, nl, pl, pt, ro, ru, sk, sl, sv, tr, uk, vi, na	Supertonic 3	GGUF F32, Stream
vibevoice	TTS, Dialogue	en, zh	VibeVoice-1.5BVibeVoice-7B	GGUF 16/Q8, Stream, LoRA
voxcpm2	TTS, Clone, Design, Ctrl	ar, da, de, el, en, es, fi, fr, he, hi, id, it, ja, km, ko, lo, ms, my, nl, no, pl, pt, ru, sv, sw, th, tl, tr, vi, zh	VoxCPM2-2B, 48 kHzVocalRender-Pro (score-conditioned singing)	GGUF 16/Q8, Stream, LoRA
voxtral_tts	TTS	en, fr, es, de, it, pt, nl, ar, hi	Voxtral-4B-TTS-2603, 20 embedded preset voices	GGUF BF16/Q8, Stream

Speech Recognition And Analysis

Family	Task	Lang	Variants	Runtime
ast_audioset	Classify	language agnostic	MIT AST AudioSet, 527 classes	GGUF F32
ced	Classify	language agnostic	CED TinyCED MiniCED SmallCED Base527 AudioSet classes	GGUF F32
audio_flamingo	ASR, Audio understanding	en	Audio Flamingo 3Audio Flamingo NextMusic Flamingo 2601SpotSound (LoRA temporal grounding)	GGUF BF16/Q8/Q4, LoRA
canary_asr	ASR, Translate	en, de, es, fr	Canary 180M Flash	GGUF F32/Q8
citrinet_asr	ASR	en	Citrinet-256	GGUF Q8
cohere_asr	ASR	en, fr, de, es, it, pt, nl, pl, el, ar, ja, zh, vi, ko	Cohere Transcribe 03-2026	GGUF BF16/Q8/Q4_0
crisperwhisper	ASR, Align	99 language codes	CrisperWhisper 2.0 Large	GGUF BF16/Q8, Stream
gigaam_asr	ASR	ru, kk, ky, uz, en	GigaAM v3 CTCGigaAM v3 RNN-TGigaAM v3 E2E CTCGigaAM v3 E2E RNN-TGigaAM Multilingual CTCGigaAM Multilingual Large CTC	GGUF F16/F32
fun_asr_nano	ASR	auto, zh, en, ja	Fun-ASR-Nano-2512	GGUF 16/Q8
higgs_audio_stt	ASR	en	Higgs Audio v3 STT	GGUF 16/Q8, Stream
hviske_asr	ASR	da	Hviske v5.3	GGUF Q8
hviske_asr_v6	ASR	da	Hviske v6	GGUF BF16 / Q8
index_echo	ASR, Translate	zh → en, es, ja	Index-Echo-S2TT 2BIndex-Echo-S2TT 9B	GGUF original/Q8_0/Q4_K
lfm_d1_omni	Classify	multilingual	d1-Omni-2B calibrated audio decisions with named yes/no, choice, score, and numeric-estimate questions	GGUF Q8_0/F16/BF16
marblenet_vad	VAD	lang agnostic	MarbleNet VAD	Bundled
micro_wake_word	Wake	model dependent	microWakeWord MixedNet TFLite models converted to GGUF	GGUF F32, Stream
ming3_audio	ASR, Understanding, TTS, Clone, S2S	zh, en	Ming 3.0 Audio Tiny	GGUF BF16/Q8_0/Q4_K
open_wake_word	Wake	en	OpenWakeWord Alexa	GGUF F32, Stream
firered_vad	VAD	lang agnostic	FireRed VAD / Stream-VAD	GGUF F32/F16/Q8, Stream
pulsevad	VAD	lang agnostic	PulseVAD 2.1K Student / 81K Teacher	GGUF F32
moonshine_asr	ASR	en	Moonshine Streaming Tiny/Small/Medium	GGUF Q8, Stream
moss_transcribe_diarize	ASR	auto, 50+ languages	MOSS-Transcribe-Diarize with speaker labels and timestamps	GGUF BF16/Q8/Q4_K, Stream
nemotron_3_diar	Diar	multilingual	NVIDIA Nemotron 3 Diarization with eight-speaker arrival-order diarization	GGUF BF16, Batch, Stream
nemotron_asr	ASR	100+ ASR prompt codes incl. auto	Nemotron 3.5 ASR Streaming 0.6B	GGUF 16/Q8, Stream
niagara_asr	ASR	en	Niagara 19M Batch EnglishNiagara 38M Batch English	GGUF F32
owsm	ASR, Translate	151 language codes	OWSM v4 Base 102MOWSM v4 Small 370MOWSM v4 Medium 1B	GGUF F32 / Q8_0 / Q4_K (Small, Medium)
owsm_ctc	ASR, Translate	151 language codes	OWSM-CTC v4 1B	GGUF F32 / Q8_0
qwen3_asr	ASR	zh, en, yue, ar, de, fr, es, pt, id, it, ko, ru, th, vi, ja, tr, hi, ms, nl, sv, da, fi, pl, cs, fil, fa, el, ro, hu, mk	Qwen3-ASR-0.6BQwen3-ASR-1.7B-hfQwenCleo-ASR (Egyptian Arabic fine-tune)	GGUF 16/Q8, Stream, LoRA
qwen3_forced_aligner	Align	zh, yue, en, de, es, fr, it, pt, ru, ko, ja	Qwen3-ForcedAligner-0.6B	GGUF 16/Q8
samsone	Audio understanding	en	SAMSONE 99MSAMSONE 134MSAMSONE 356M	GGUF BF16/Q8
silero_vad	VAD	lang agnostic	Silero VAD	Bundled, Stream
sherpa_kws	Wake	model dependent	sherpa-onnx streaming Zipformer keyword spotting	GGUF F32, Stream
smart_turn	Turn detection	auto	Smart Turn v3.2	GGUF F32
sortformer_diar	Diar	en	Sortformer-4spk-v1	-
vibevoice_asr	ASR	auto	VibeVoice ASR	GGUF 16/Q8
vibevoice_asr_streaming	ASR	en, zh, es, pt, de, ja, ko, fr, ru, it	VibeVoice ASR Streaming 7B/1.5B with persistent decoder state and speaker turns	GGUF BF16/Q8/Q4, Stream
voxtral_realtime	ASR	auto	Voxtral-Mini-4B-Realtime-2602	GGUF 16/Q8/Q4, Stream

Audio Conversion And Processing

Family	Task	Lang	Variants	Runtime
apollo	S2S	lang agnostic	Apollo music restoration	GGUF F32
audiosr	S2S	lang agnostic	AudioSR Basic audio super-resolution package	GGUF F32
bs_roformer	Sep	lang agnostic	BS-RoFormer vocal separation checkpoints	GGUF Q8
builtin_audio_utils	S2S	lang agnostic	RNNoiseDeepFilterNet2ZipEnhancerGTCRN StreamingGTCRN DNS3GTCRN VCTKFlashSR	Safetensors
controlfoley	SFX	auto	ControlFoley 44 kHz multimodal Foley generation from text, video, and reference audio conditioning	GGUF F32/Q8
htdemucs	Sep	lang agnostic	HTDemucsHTDemucs_ftHTDemucs_6stems	GGUF 16/Q8
meanvc2	VC	lang agnostic	MeanVC2 120 ms/40 ms zero-shot voice conversion	GGUF F32/Q4, Stream
mel_band_roformer	Sep	lang agnostic	Mel-Band RoFormer MLX vocal separation variants	GGUF 16/Q8
miocodec	Codec, VC	lang agnostic	MioCodec v2, 25 Hz, 44.1 kHz	GGUF 16/Q8
mossformer2	Sep	lang agnostic	MossFormer2 SS 16K two-speaker separation	GGUF F32
muscriptor	MIDI	music	MuScriptor Small/Medium/Large audio-to-symbolic transcription	GGUF F32, Stream
rvc	VC	lang agnostic	RVC F16 GGUF with packaged v1/v2 voices and optional retrieval blending	GGUF 16
sam_audio	S2S	lang agnostic	SAM Audio SmallSAM Audio BaseSAM Audio Large	GGUF F32/BF16/Q8
seed_vc	VC	lang agnostic	SeedVC XLS-R + HiFTSeedVC Whisper-small + BigVGAN	GGUF 16/Q8
sheetsage2	MIDI	music	SheetSage2 audio-to-ABC score transcription	GGUF orig
sidon	S2S	lang agnostic	Sidon v0.1 single-speaker speech restoration	GGUF F32
tf_gridnet	Sep	lang agnostic	TF-GridNet WSJ0-2mix two-speaker separation	GGUF F32
tone_color_vc	VC	lang agnostic	Tone Color VC	GGUF F32/F16
universr	S2S	lang agnostic	UniverSR Audio and Speech super-resolution	GGUF F32

Music, Media, And Editing

Family	Task	Lang	Variants	Runtime
ace_step	Music, Edit	50+ langs	ACE-Step 1.5 TurboACE-Step 1.5 BaseACE-Step 1.5 XL TurboACE-Step 1.5 XL SFT	GGUF 16, LoRA
heartmula	Music	zh, en, ja, ko, es	HeartMuLa-oss-3B with HeartCodec-oss	GGUF 16/Q8
magenta_realtime_2	Music	music	Magenta RealTime 2 Small (230M)Magenta RealTime 2 Base (2.4B)	GGUF F16/Q8_0, Stream
midashenglm_gen	Music, SFX	auto	MiDashengLM-Gen structured-prompt generation for speech, music, sound effects, and ambience	GGUF F32/Q8
minimax_h3	Video, Music, TTS/Dialogue	auto	MiniMax-H3 Q4_K with optional INT8 ConvRot DiT	GGUF Q4/INT8
minimax_music3	Music	auto	MiniMax Music 3 text-to-music generation with lyrics conditioning	GGUF Q4/Q8
moss_soundeffect	SFX	en, zh	MOSS-SoundEffect v2 text-conditioned sound effects	GGUF BF16/Q8
mulacover	Music	multilingual	MuLaCover cover-song generation from reference audio or melody/chord MIDI, with new lyrics and style	GGUF 16/Q8
soulx_flashhead	Audio-to-video	lang agnostic	SoulX-FlashHead Lite audio-driven portrait video generation	GGUF Q8/BF16, Stream
stable_audio	Music, SFX, Edit	en	Stable Audio 3 Small MusicStable Audio 3 Small SFXStable Audio 3 Medium	GGUF 16/Q8
vevo2	TTS, Music, VC, Edit	en, zh	Vevo2 with Qwen2.5-0.5B AR model	GGUF 16
yue2	Music	en	YuE2-3B lyrics-conditioned song generation with optional ABC score planning and conditioning	GGUF BF16/Q8/Q4, LoRA

Some model families in the supported table started as outside contributions before being promoted into the core release surface. Thanks to Mirek @mirek190 for BS-RoFormer, @justinjohn0306 for MOSS-TTS-Local, and @LauraGPT from the official FunASR team for Fun-ASR-Nano.

Community Models

Community model ports live under community_models to make the ownership boundary clear while keeping them available through the normal audio.cpp CLI and server paths. Some community-contributed models graduate into the core model tree when they become part of the main release surface. Huge thanks to the contributors who bring these models in, test them, and keep pushing the framework into new territory. See CONTRIBUTING.md for community-model expectations.

Family	Task	Lang	Runtime	Contributor	What They Added
audio8_asr	ASR	en, zh, yue, ja, ko, fr, de	GGUF Q8, Safetensors	@gqf2008	Audio8-ASR-0.1B compact multilingual autoregressive ASR reusing the Qwen3-ASR encoder with an MLP-tower adapter and an 8-layer Qwen2-style decoder (CC-BY-NC, local conversion only)
audio8_tts	TTS, Clone	auto, yue, zh, nl, en, fr, de, it, ja, ko, pl, es	GGUF Q8, Stream	@jasonchen31	Audio8 TTS Preview 0.6B DualAR multilingual TTS and zero-shot voice cloning with a Qwen backbone and neural codec
auk	TTS, audio editing	auto	GGUF F32/F16/Q8	@0xShug0	AuK Base and AuK-Flash instruction-guided speech generation and editing with separate generator, Qwen conditioner, and VAE components
chatterbox_turbo	TTS (testing)	en	GGUF 16/Q8	@pannagaps	Chatterbox Turbo distilled 350M GPT2 T3 backbone + 2-step meanflow S3Gen decoder; built-in voice
confucius4_r2t2	ASR	30 languages	Safetensors, GGUF F16/Q8/Q4, Stream	@davidxifeng	Confucius4-R2T2 streaming ASR with context and hotword prompts
earshot_vad	VAD	lang agnostic	Bundled, Stream	@Etherll	Earshot VAD native C++ port with offline and streaming speech detection
echo_tts	TTS, Clone	en	GGUF 16/Q8	@5uck1ess	Echo-TTS 44.1 kHz zero-shot voice cloning with EchoDiT latents and Fish S1-DAC decoding
f5_tts	TTS, Clone	en, ar (Habibi)	GGUF, LoRA	@tareko	F5-TTS flow-matching DiT synthesis and voice cloning, with Habibi Arabic aliases habibi/habibi_tts
fish_drama3	TTS, Clone, Ctrl, Edit	auto	Mixed Q2_K/Q3_K/Q4_K GGUF, IQ3_XXS (small), Stream	@JarodMica (base), @Th-Underscore (small)	Experimental Fish Audio Drama 3 port and Drama3-small variant support
glm_tts	TTS, Clone	zh, en	GGUF	Mirek @mirek190	GLM-TTS zero-shot synthesis and voice cloning support
granite5asr	ASR	en, fr, de, es, pt, it, ja	GGUF Q8	@ampersandru	IBM Granite Speech 5.0 470M TurboCTC ultra-fast Conformer-CTC ASR with Shaw relative positional embeddings and ByteLevel BPE
haas_asr	ASR	ja	GGUF F16/Q8, Stream	Third Intelligence	Haas-ja-v1.0, sharing Parakeet’s encoder and TDT decoder with a KDA predictor
inflect_v2	TTS	en	GGUF FP32	Jan @JanWerder	Inflect Micro v2 and Nano v2 native offline synthesis
kitten_tts	TTS	en	GGUF FP32	Community	KittenTTS Mini 0.8 native 80M-parameter synthesis with eight built-in voices and the shared eSpeak-ng frontend
kitten_tts2	TTS, Clone	en, ar, zh, fr, de, hi, it, pt, ru, es, nl, id, ja, ko, pl, tr	GGUF Q8/Q4	@dignome	Kitten TTS 2 native Qwen3/S3 synthesis, multilingual presets, and reference-audio cloning; ready-to-run GGUF
kroko_asr	ASR	de, en, es, fr, it, he, nl, pt, sv, tr	Safetensors, GGUF Q8	Mirek @mirek190	Kroko Community ASR native offline/streaming Zipformer2/RNN-T transcription with word timestamps
latentsync	Video lip sync	lang agnostic	GGUF F32/F16/Q8, experimental	@0xShug0	LatentSync 1.6 audio-conditioned lip synchronization with optional DeepCache and reduced-memory execution modes
lfm2_audio	ASR, TTS, S2S	en, ja	GGUF F32/16/Q8/Q4, Stream	@ykhrustalev / Liquid AI	LFM2.5-Audio English and Japanese transcription, speech and spoken chat from Liquid AI's published GGUFs: LFM2 hybrid short-conv/attention backbone with a FastConformer encoder for audio input, and a depthformer and LFM2 detokenizer for 24 kHz speech in four built-in English voices; spoken questions get replies of interleaved text and speech; offline or streamed frame by frame
lfm3_audio	ASR, TTS, S2S	en, ar, da, de, el, es, fi, fr, he, it, ja, ko, nl, pl, pt, sv, tr, zh	GGUF BF16/16/Q8/Q4, Safetensors, Stream, experimental	Community	LFM3-Audio transcription, spoken-question chat (with optional thinking), speech in 38 built-in voices across 19 languages and spoken chat from Liquid AI's published LFM3-Audio-2B GGUFs: LFM2 encoder for audio input, LFM3 hybrid short-conv/attention backbone, a bidirectional MaskGIT depth decoder (12 codebooks, 12.5 Hz) and the LiquidCodec decoder for 24 kHz speech; spoken questions get replies of interleaved text and speech, with client-held multi-turn history; TTS and S2S offline or streamed frame by frame
liveavatar	Audio-to-video	auto	GGUF NVFP4	@0xShug0	LiveAvatar audio-to-video avatar generation with Wan2.2 S2V and an optional low-VRAM denoiser path
minimax_h3	Video, Music, TTS/Dialogue	auto	GGUF Q4/INT8	@0xShug0	MiniMax-H3 text-to-audio/video generation with Q4_K and optional INT8 ConvRot DiT
minimax_music3	Music	auto	GGUF Q4/Q8	@0xShug0, @JoeMattie	MiniMax Music 3 text-to-music generation with lyrics conditioning
mira_tts	TTS, Clone	en	Local conversion	Mirek @mirek190	MiraTTS experimental native Qwen2 + ECAPA/Perceiver zero-shot voice cloning with progressive segment streaming (CC-BY-NC-SA-4.0 weights)
mms_forced_aligner	Align	nl (nld), en (eng); pre-romanized Latin	Safetensors, GGUF 16/Q8	@LysanderdeJong	MMS-300M-1130 Forced Aligner word-timestamp alignment from a wav2vec2 CTC checkpoint (safetensors or local GGUF)
nemotron_transcribe	ASR, translation, audio Q&A	auto	IQ4_NL / Q4_K / Q3_K / Q2_K experts	@JarodMica	NVIDIA Nemotron 3.5 Transcribe 30B-A4B experimental offline speech language model with CTC word timestamps and speaker labels
moss_tts_v15	TTS, Clone	en, zh	GGUF	Chris @christopherthompson81	MOSS-TTS-v1.5 8B delay-pattern zero-shot voice cloning; voice-attribute instructions are followed only loosely, see the limitation note
moss_ttsd	TTS, Clone, Dialogue	en, zh	GGUF BF16/Q8/Q4	Chris @christopherthompson81	MOSS-TTSD 8B dialogue TTS with speaker-tagged conversations and per-speaker voice cloning
moss_voicegen	Voice Design	en, zh	GGUF	Joost @jrohde	MOSS-VoiceGenerator speech in a voice designed from a written instruction
outetts	TTS, Clone	en, ar, zh, nl, fr, de, it, ja, ko, lt, ru, es, pt, be, bn, ka, hu, lv, fa, pl, sw, ta, uk	GGUF, LoRA	Mirek @mirek190	Llama-OuteTTS-1.0-1B TTS and voice cloning support
parakeet_tdt	ASR	auto, bg, cs, da, de, el, en, es, et, fi, fr, hr, hu, it, lt, lv, mt, nl, pl, pt, ro, ru, sk, sl, sv, uk	GGUF F32/16/Q8, Stream	@dleiferives	Parakeet-TDT 0.6B v3 offline, long-form, and buffered-streaming ASR support, plus Orukeet r3 weights
phonon2_asr	ASR	en	GGUF F32/16/BF16/Q8, Stream	Mirek @mirek190	Phonon-2, native hotwords and phrase-based streaming on the shared Parakeet implementation
piper_tts	TTS	en	GGUF FP32	Community	Piper TTS native VITS synthesis for the Lessac medium voice with the shared eSpeak-ng frontend
realtime_reuse	S2S	lang agnostic	GGUF F32, Stream	@0xShug0	Real-time RE-USE causal speech restoration with 30 exit-layer/lookahead configurations
reuse	S2S	lang agnostic	GGUF F32/F16/Q8_0	@0xShug0	RE-USE speech restoration
sanotts	TTS	en, vi, id, cs, de, es, fr, it, pt, ro, ru, tr, ne, hi	GGUF FP32	Ashish @voidash	sanoTTS voice family eighteen voices from 294k to 2.27M parameters, native offline synthesis
sense_asr	ASR	auto, zh, en, yue, ja, ko, pt, ru, es, it, fr, de, nl, pl, tr, ar, hi, vi, th, id, ms, fa, nospeech	GGUF Q8, Stream	Jason Chen @jasonchen31, @LauraGPT / FunASR	SenseVoice-Small offline/streaming SAN-M + CTC transcription with event/emotion/language tags, ITN and word timestamps
soprano_tts	TTS	en	GGUF Q8, Stream	@drzsdrtfg	Soprano-1.1-80M ultra-lightweight TTS with Qwen3 LM + Vocos decoder
sopro_tts	TTS, Clone	en, pt, fr, de	Safetensors, GGUF F16, Stream	@Brioch, @samuel-vitorino (model author)	Sopro V2 Turbo updated for Sopro 2.2.0: 120M zero-shot voice cloning, chunked streaming, and optional voice caching
sortformer_diar_v2	Diar	multilingual	GGUF F32/mixed F16, Stream	@LysanderdeJong	NVIDIA Sortformer v2.1 four-speaker streaming diarization; local conversion only pending redistribution approval
underdog_bark3	ASR	en	GGUF F32/16/BF16/Q8, Stream	Mirek @mirek190	Underdog Bark 3 Standard/Plus/Pro, sharing the Parakeet implementation
vibeasr	ASR	en	GGUF I8_S + I2_S	@XsquirrelC / VibeASR	VibeASR fully quantized port of VibeASR.cpp: VibeVoice acoustic/semantic tokenizers on INT8 weights and INT8 activations, feeding a ternary BitNet Qwen2 decoder. Offline, CPU only
vieneu_v3_turbo	TTS, Clone	vi, en	GGUF	Bảo @pnnbao97 (model author); first port by Phuoc @phuocnguyen90	VieNeu-TTS v3 Turbo 48 kHz Vietnamese/English TTS and voice cloning (formerly vietneu_tts)
voxcpm1	TTS, Clone	zh, en, ja, ko	GGUF Q8, Stream, LoRA	@jasonchen31	VoxCPM1 tokenizer-free 0.5B TTS with 16 kHz output, streaming, and continuation-mode voice cloning
whistle_asr	ASR	en, de, fr, es, it, nl, pl	GGUF F32, Safetensors F32	@bumpyclock	Whistle ASR experimental offline transcription with a GGML encoder on CPU, Metal, CUDA, or Vulkan and a host decoder; 16 kHz mono, at most 30 seconds
xtts_v2	TTS, Clone	en, es, fr, de, it, pt, pl, tr, ru, nl, cs, ar, zh-cn, hu, ko, ja, hi	GGUF original/Q8_0/Q4_K_M	@whoreson	Coqui XTTS v2 multilingual speech generation with a reference voice and cached speaker conditioning
zipvoice	TTS, Clone	zh, en	GGUF F32, Stream	Community	ZipVoice k2-fsa TTSZipformer flow-matching zero-shot voice cloning with Vocos vocoder; jieba + pypinyin Chinese frontend; chunk streaming

Prebuilt Binaries

Official packages for every release are attached to the Releases page:

Platform	Backends
Windows x64	CPU, Vulkan, CUDA (12.4 / 13.3), HIP/ROCm (test)
Ubuntu x64	CPU, Vulkan, HIP/ROCm (test)
macOS (arm64 / x64)	Metal

The Windows CUDA packages ship the CUDA runtime in a separate cudart archive. Extract it next to the binaries so the CUDA backend can load ggml-cuda.dll and the CUDA runtime DLLs (cudart, cuBLAS, cuFFT).

• HIP/ROCm (AMD GPUs): Official Windows and Ubuntu test builds are available on the Releases page. GPU compatibility and model coverage are still being validated; please report issues with your GPU, driver version, and reproduction command.
• Older community Windows HIP packages: @IIIIIllllIIIIIlllll's fork Releases provides v0.9.0 packages with the ROCm runtime bundled, in ROCm 6.4 and 7.1 tracks. See docs/build/windows-hip-distribution.md for packaging details.

Homebrew Install

On macOS, audio.cpp can be installed from the Homebrew tap:

brew tap 0xShug0/audio-cpp
brew trust 0xShug0/audio-cpp
brew install audio-cpp

WebUI

audiocpp_server includes an embedded SvelteKit/TypeScript WebUI for running local TTS, cloning, ASR, generation, conversion, separation, VAD, diarization, and alignment workflows. The production UI is compiled into the server binary, so using it requires neither Python nor separate frontend files:

audiocpp_server --ui --ui-management --backend cuda

Open http://127.0.0.1:8080. --ui-management enables catalog browsing, downloads, temporary uploads, and dynamic model switching, so this is the easiest way to try audio.cpp without writing a server config first. For a locked-down config-driven instance, start with --ui and a server config instead; in that mode the UI only offers models declared by the config unless management is enabled.

For Docker UI downloads, mount a writable models directory at /app/models and use --ui-management. See docker.md.

The native UI also exposes background model download/preparation, long-text split-and-merge synthesis, a browser-local saved voice library, microphone recording, and near-live ASR input. Some model preparation jobs invoke the repository's Python model manager because those packages require Hugging Face download or checkpoint conversion; model inference and the embedded UI remain Python-free. See webui/README.md for launch commands, model notes, and frontend development instructions.

Docker

Docker CUDA and CPU images are available for both CLI and server use. See docker.md for available images, build commands and working Docker examples.

Model Manager and GGUF Downloads

Use tools/model_manager_v2.py for normal model downloads. It reads model_specs/*.json and installs the default package for each family, preferring ready-to-use GGUF packages when they are available.

Native builds configured with -DAUDIOCPP_BUILD_NATIVE_MODEL_MANAGER=ON also provide audiocpp_model_manager, a standalone headless frontend over the same reusable C++ package-management library used by the server and embedded WebUI. It supports list, info, install, clean, and remove without starting audiocpp_server; this is the preferred native path for CLI, Docker, remote provisioning, and other scripted environments. The Python v2 manager remains available as an alternative during migration. The native path uses bundled TLS by default and does not require libcurl.

audiocpp_model_manager list
audiocpp_model_manager install qwen3_asr_0_6b_q8_0 --models-dir models

The old safetensors/converter catalog has been renamed to tools/model_manager_deprecated.py. Use it only for legacy model layouts that have not moved to spec-backed GGUF packages yet.

GGUF downloads:

• Official model packages: audio-cpp on Hugging Face
• Community-hosted model packages: see the individual model guides linked in the supported model tables for repositories and download instructions.

See the Model Manager guide for model-manager usage and package notes.

Build yourself

OS	Requirements
Linux	GCC 13 or newer, CMake, plus the backend toolchain for the build you want: NVIDIA CUDA Toolkit for CUDA, Vulkan SDK for Vulkan, ROCm for HIP
Windows	Visual Studio Build Tools 2022 or newer with C++ desktop workload, MSVC x64 compiler, Windows SDK, CMake, Ninja, MSVC OpenMP components; official NVIDIA CUDA Toolkit for CUDA builds, Vulkan SDK for Vulkan builds, AMD HIP SDK for HIP builds
macOS	Xcode or Xcode Command Line Tools, plus CMake. Metal builds also require the Metal compiler available through xcrun

Composite builds let you compile only the model families you need. full is the default and is what release/Docker builds should use. custom registers only the requested loaders while still linking required internal dependencies; core builds the runtime without the optional model-family set.

The helper scripts expose this as --model-set and --models on Linux/macOS, and -ModelSet and -Models on Windows:

scripts/build_linux.sh --backend cuda --model-set custom --models qwen3_tts,pocket_tts,qwen3_asr --target audiocpp_cli

Direct CMake builds use the same underlying variables:

cmake -S . -B build/debug -DCMAKE_BUILD_TYPE=Debug -DAUDIOCPP_MODEL_SET=custom -DAUDIOCPP_MODELS=qwen3_tts,pocket_tts,qwen3_asr
cmake --build build/debug --target audiocpp_cli -j 8

Use the Linux helper script for CPU, CUDA, Vulkan, or HIP builds:

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 hip --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend cpu --target audiocpp_cli --target audiocpp_server

The script writes to aligned build directories such as build/linux-cuda-release, build/linux-vulkan-release, build/linux-hip-release, and build/linux-cpu-release.

Without --cuda-arch, CUDA builds use the portable arch list (works on many GPUs, slower to build). For a faster build targeting only the local GPU, pass --cuda-arch native (CMake >= 3.24; on older CMake this falls back to the portable list) or an explicit arch list:

scripts/build_linux.sh --backend cuda --cuda-arch native --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend cuda --cuda-arch "86;89" --target audiocpp_cli --target audiocpp_server

Nothing caches compiler output by default. --ccache wires ccache up as the compiler launcher, which leaves a cold build about as slow as it already is and makes a rebuild roughly 14x faster. A CMAKE_<LANG>_COMPILER_LAUNCHER already set in the environment is left alone.

scripts/build_linux.sh --backend cuda --cuda-arch native --ccache --target audiocpp_cli

Composite examples:

scripts/build_linux.sh --backend cuda --model-set full --target audiocpp_cli
scripts/build_linux.sh --backend cuda --model-set custom --models qwen3_tts,pocket_tts,qwen3_asr --target audiocpp_cli
scripts/build_linux.sh --backend cpu --model-set core --target audiocpp_cli

For portable CPU kernels on machines where native ISA flags are not suitable:

scripts/build_linux.sh --backend cuda --native-cpu OFF --target audiocpp_cli --target audiocpp_server

For deployment builds with compiled package specs:

scripts/build_linux.sh --backend cuda --deployment-build --target audiocpp_cli --target audiocpp_server

For native WebUI model downloads, enable the native model manager:

scripts/build_linux.sh --backend cuda --native-model-manager --target audiocpp_server

For direct CMake commands, see docs/build/linux.md. For the opt-in Rockchip NPU (RKNN) decoder path, see docs/build/rockchip-npu.md.

Use the Windows PowerShell build script:

powershell.exe -NoProfile -ExecutionPolicy Bypass -File .\scripts\build_windows.ps1

Common presets:

.\scripts\build_windows.ps1 -Preset windows-cuda-release -Target audiocpp_cli
.\scripts\build_windows.ps1 -Preset windows-vulkan-release -Target audiocpp_cli
.\scripts\build_windows.ps1 -Preset windows-cpu-release -Target audiocpp_cli
.\scripts\build_windows.ps1 -Target audiocpp_server -Jobs 16
.\scripts\build_windows.ps1 -Preset windows-cuda-release -ModelSet custom -Models "qwen3_tts,pocket_tts,qwen3_asr" -Target audiocpp_cli

From cmd.exe, use the wrapper:

scripts\build_windows.cmd

For deployment builds with compiled package specs:

.\scripts\build_windows.ps1 -DeploymentBuild -Target audiocpp_cli

For native WebUI model downloads, enable the native model manager:

.\scripts\build_windows.ps1 -NativeModelManager -Target audiocpp_server

For requirements, CPU profiles, CUDA packaging, and release zips, see docs/build/windows.md.

CPU-only Build

Apple builds enable Metal by default. To build a CPU-only binary, disable Metal explicitly:

cmake -S . -B build/macos-cpu-release \
  -DCMAKE_BUILD_TYPE=Release \
  -DENGINE_ENABLE_CUDA=OFF \
  -DENGINE_ENABLE_VULKAN=OFF \
  -DENGINE_ENABLE_METAL=OFF \
  -DENGINE_ENABLE_OPENMP=OFF \
  -DGGML_OPENMP=OFF
cmake --build build/macos-cpu-release \
  --parallel "$(sysctl -n hw.logicalcpu)" \
  --target audiocpp_cli audiocpp_server audiocpp_gguf

Confirm that the resulting CLI sees the host CPU backend:

build/macos-cpu-release/bin/audiocpp_cli --list-devices

Metal Build

On macOS, use the Metal helper script to build against ggml's Metal backend:

scripts/build_metal.sh --target audiocpp_cli

The script configures build/macos-metal-release by default, enables ENGINE_ENABLE_METAL=ON, disables CUDA and Vulkan, embeds the Metal shader library, and builds static libraries plus the requested target.

Useful variants:

scripts/build_metal.sh --target audiocpp_server
scripts/build_metal.sh --build-type Release --archs arm64 --target audiocpp_cli
scripts/build_metal.sh --model-set custom --models qwen3_tts,pocket_tts --target audiocpp_cli
scripts/build_metal.sh --with-tests --target audio_dsp_test
scripts/build_metal.sh --openmp auto --target audiocpp_cli
scripts/build_metal.sh --native-cpu OFF --target audiocpp_cli
scripts/build_metal.sh --deployment-build --target audiocpp_cli

The built CLI is written to:

build/macos-metal-release/bin/audiocpp_cli

On Linux and Windows, HIP builds compile ggml's CUDA backend sources as HIP code for AMD GPUs. ENGINE_ENABLE_HIP and ENGINE_ENABLE_CUDA are mutually exclusive — configure with exactly one of them.

Linux (the helper script auto-detects ROCm via ROCM_PATH/HIP_PATH/hipconfig and local GPU targets via amdgpu-arch, falling back to rocminfo; pass --gpu-targets to build for other architectures, or when no AMD GPU is visible, e.g. in a VM or container):

scripts/build_linux.sh --backend hip --target audiocpp_cli --target audiocpp_server
scripts/build_linux.sh --backend hip --gpu-targets "gfx1100;gfx1103" --target audiocpp_cli

Direct CMake:

cmake -S . -B build_hip \
  -DENGINE_ENABLE_HIP=ON \
  -DGPU_TARGETS=gfx1151 \
  -DCMAKE_C_COMPILER="$(hipconfig -l)/clang" \
  -DCMAKE_CXX_COMPILER="$(hipconfig -l)/clang++" \
  -DCMAKE_BUILD_TYPE=Release
cmake --build build_hip -j$(nproc)

Windows (the helper script auto-detects ROCm, GPU targets, cmake, and ninja):

powershell -ExecutionPolicy Bypass -File scripts\build_windows_hip.ps1

Run with --backend hip (rocm is accepted as an alias). For GPU target selection, hipBLASLt GEMM notes, iGPU tuning, and known limitations, see docs/build/HIP.md.

Native WebGPU builds use Dawn. Enable the backend explicitly and point CMake to the directory containing DawnConfig.cmake:

cmake -S . -B build/debug \
  -DCMAKE_BUILD_TYPE=Debug \
  -DENGINE_ENABLE_WEBGPU=ON \
  -DDawn_DIR=/path/to/dawn/lib/cmake/Dawn
cmake --build build/debug --target audiocpp_cli -j 8

Run with --backend webgpu. Browser builds use Emscripten to run the C++ CLI and load GGUF models inside a Web Worker, with playable/downloadable outputs. They require a browser with WebGPU, shader-f16, and JSPI support.

For native and browser build commands, WASM32/WASM64 memory limits, and Kokoro/eSpeak setup, see docs/build/webgpu.md.

For Nix and NixOS builds, see the build guide.

Optional server adapters are built from the external/audio.cpp-server-frontends submodule. They are separate from the embedded WebUI. From the repository root, fetch the package (using HTTPS here so no GitHub SSH key is required):

git -c 'url.https://github.com/[email protected]:' submodule update --init external/audio.cpp-server-frontends

For example, build a server with audio decoding and MP3 encoding adapters:

cmake -S . -B build/debug \
  -DCMAKE_BUILD_TYPE=Debug \
  -DAUDIOCPP_BUILD_SERVER_FRONTENDS=ON \
  -DAUDIOCPP_SERVER_FRONTENDS_DIR=external/audio.cpp-server-frontends \
  -DAUDIOCPP_SERVER_FRONTEND_MODULES="audio_decode;mp3_encode"
cmake --build build/debug --target audiocpp_server --parallel 8

The example uses the default backend settings of a fresh build directory; add the backend flags from the platform build instructions when needed. Module-specific dependencies and available adapters are documented in the frontend package. See the server guide for configuration and usage.

Option	Meaning	Default
ENGINE_ENABLE_CUDA	Enable the ggml CUDA backend. Required for --backend cuda.	OFF
ENGINE_ENABLE_HIP	Enable the ggml HIP backend (AMD GPUs). Required for --backend hip; mutually exclusive with ENGINE_ENABLE_CUDA.	OFF
ENGINE_ENABLE_VULKAN	Enable the ggml Vulkan backend. Required for --backend vulkan.	OFF
ENGINE_ENABLE_METAL	Enable the ggml Metal backend. Required for --backend metal.	OFF on most platforms, ON on Apple
ENGINE_ENABLE_LLAMAFILE	Enable llamafile SGEMM support in ggml CPU builds.	ON
ENGINE_ENABLE_CUDA_GRAPHS	Enable ggml CUDA graphs support when CUDA is enabled.	ON
ENGINE_ENABLE_NATIVE_CPU	Build ggml CPU kernels with native host ISA flags such as -march=native. Disable this for portable CPU kernels or toolchains that reject generated CPU instructions.	ON
ENGINE_ENABLE_OPENMP	Enable OpenMP for host-side parallel work.	ON
ENGINE_BUILD_EXAMPLES	Build example binaries.	OFF
AUDIOCPP_BUILD_SERVER_FRONTENDS	Enable optional in-process server frontend adapters, separate from the embedded WebUI.	OFF
AUDIOCPP_SERVER_FRONTEND_MODULES	Semicolon-separated adapter modules to build. Requires AUDIOCPP_BUILD_SERVER_FRONTENDS=ON and an external frontend package.	empty
AUDIOCPP_SERVER_FRONTENDS_DIR	External frontend package directory containing server_frontends.cmake, which registers the requested modules.	empty
ENGINE_BUILD_TESTS	Build framework unit tests.	OFF
ENGINE_BUILD_EXTENDED_TESTS	Build extended non-model tests and probes, such as server/app/package-manager checks and backend graph tests.	OFF
ENGINE_BUILD_MODEL_TESTS	Build model-specific tests and probes.	OFF
ENGINE_BUILD_WARMBENCH	Build warmbench helper binaries.	OFF
AUDIOCPP_BUILD_NATIVE_MODEL_MANAGER	Build the standalone native model manager and enable server-side WebUI downloads/install management. This opt-in feature builds the HTTP/TLS dependency.	OFF
AUDIOCPP_USE_SYSTEM_OPENSSL	Use system OpenSSL instead of bundled BoringSSL when native model management is enabled.	OFF
AUDIOCPP_DEPLOYMENT_BUILD	Compile package specs into CLI/server binaries for standalone GGUF and package-spec fallback loading. Script builds expose this as --deployment-build on Linux/macOS and -DeploymentBuild on Windows.	OFF
Native model manager script flag	Build scripts expose AUDIOCPP_BUILD_NATIVE_MODEL_MANAGER as --native-model-manager on Linux/macOS and -NativeModelManager on Windows. --system-openssl / -SystemOpenSsl and --boringssl-archive / -BoringSslArchive configure its TLS backend.	disabled
AUDIOCPP_MODEL_SET	Model composite to build: full, core, or custom. Script builds expose this as --model-set on Linux/macOS and -ModelSet on Windows.	full
AUDIOCPP_MODELS	Comma or semicolon separated model target names when AUDIOCPP_MODEL_SET=custom, such as qwen3_tts,pocket_tts,qwen3_asr. Script builds expose this as --models on Linux/macOS and -Models on Windows.	empty

Usage

[!IMPORTANT] Please read your model's documentation: open the model-docs index, choose your task category, then find your model. Also read the dedicated CLI usage guide and server guide. The examples below are quick starts; model requirements, supported options, and setup details vary.

CLI

The main CLI binary is:

build/bin/audiocpp_cli

High-level command shape:

audiocpp_cli --task <task> --model <path> --family <family> [--backend <backend>] [--mode <mode>] [options]

Always specify --family for model commands, including model-specific help. Treat it as required; omission is supported only for backward compatibility. Automatic loader discovery without it can substantially slow startup and --model ... --help.

Core selectors:

• --task vad|asr|diar|sep|gen|tts|clon|vc|s2s|align|vdes|spk|svc|midi|wake|cls
• --model <path>
• --family <name> explicitly selects the model family
• --backend cpu|cuda|vulkan|metal|best
• --mode offline|streaming; streaming is available for models whose docs list streaming support

Common interface options:

• --load-option key=value passes model-load options, such as PocketTTS language selection
• --session-option key=value passes session/runtime options, such as backend-specific weight controls
• --request-option key=value passes per-request model options
• --config <id> selects a discovered config asset
• --weight <id> selects a discovered weight asset
• --device <n> selects the backend device
• --threads <n> sets backend and OpenMP worker threads

Examples:

Text-to-speech:

build/bin/audiocpp_cli \
  --task tts \
  --family pocket_tts \
  --model /path/to/model \
  --backend cuda \
  --text "audio.cpp is running PocketTTS locally." \
  --voice-ref assets/resources/sample.wav \
  --out build/out/pocket_tts.wav

PocketTTS with another language and a built-in voice:

build/bin/audiocpp_cli \
  --task tts \
  --family pocket_tts \
  --model /path/to/models/pocket-tts \
  --backend cuda \
  --load-option language=spanish \
  --text "Hola, esta es una prueba corta de Pocket TTS." \
  --voice-id alba \
  --out build/out/pocket_tts_spanish.wav

ASR:

build/bin/audiocpp_cli \
  --task asr \
  --family qwen3_asr \
  --model /path/to/model \
  --backend cuda \
  --audio assets/resources/sample_16k.wav

Voice conversion:

build/bin/audiocpp_cli \
  --task vc \
  --family seed_vc \
  --model /path/to/model \
  --backend cuda \
  --audio assets/resources/a.wav \
  --voice-ref assets/resources/b.wav \
  --out build/out/seed_vc.wav

Useful CLI features:

• --help with --task shows task-oriented help
• --help with --model <path> and --family <family> shows model-owned request, session, and load options
• --inspect prints discovered configs, weights, and capabilities
• --list-loaders prints registered model families (--json for the machine-readable contract)
• --list-devices prints backend devices and the --backend/--device values that select them (--json adds PCI ids and free/total memory)
• python tools/model_manager_v2.py list --json prints installable packages from model_specs/*.json
• --batch-text-file <txt> runs one offline request per non-empty line
• --batch-text-dir <dir> runs one offline request per .txt, .md, or .json file, normalizing each file as one paragraph
• --batch-audio-dir <dir> runs one