TTS Models

Model	Family	Task(s)	Quick Start
Qwen3 TTS	qwen3_tts	tts, vdes	Qwen3 TTS
Chatterbox	chatterbox	clon, vc	Chatterbox
Confucius4-TTS	confucius4_tts	clon	Confucius4-TTS
DramaBox	dramabox	tts, clon	DramaBox
DotTTS	dots_tts	tts, clon	DotTTS
MioTTS	miotts	tts	MioTTS
MOSS-TTS-Local	moss_tts_local	tts, clon	MOSS-TTS-Local
MOSS-TTS-Nano	moss_tts_nano	tts, clon	MOSS-TTS-Nano
MiniMax-H3	minimax_h3	gen dialogue audio	MiniMax-H3
NeuTTS	neutts	tts	NeuTTS
OmniVoice	omnivoice	tts	OmniVoice, full guide
PocketTTS	pocket_tts	tts	PocketTTS
VoxCPM2	voxcpm2	tts, vdes	VoxCPM2
Higgs Audio v3 TTS	higgs_audio_tts	tts	Higgs Audio v3 TTS
Fish Audio S2 Pro	fish_audio	tts	Fish Audio S2 Pro
IndexTTS2	index_tts2	tts	IndexTTS2
IndexTTS2.5	index_tts2 (variant 2.5)	tts	IndexTTS2.5
Irodori-TTS	irodori_tts	tts, vdes	Irodori-TTS
GLM-TTS	glm_tts	tts, clon	GLM-TTS
Inflect Micro v2	inflect_v2	tts	Inflect v2
OuteTTS	outetts	tts, clon	OuteTTS
Supertonic	supertonic	tts	Supertonic
VibeVoice	vibevoice	tts	VibeVoice

This page covers speech TTS-style families. MiniMax-H3 appears here for prompt-driven dialogue audio, but it uses the generation route (--task gen) rather than the normal speech route (--task tts). Detailed route manuals live under docs/models/ or docs/community_models/ when a model needs more space.

Common CLI shape:

audiocpp_cli --task <task> --family <family> --model <model-dir> --backend cuda ...

Common options:

Option	Meaning
--text	Text, prompt, lyrics, or multi-speaker script, depending on the model.
--voice-ref	Reference voice WAV for models that support cloning.
--reference-text	Transcript or prompt text for models that use explicit reference transcripts.
--voice-id	Built-in voice id for models with packaged voices.
--language	Model language code when the model requires one.
--text-chunk-size	Long-form chunk budget in characters. Each model has its own default.
--seed	Optional fixed seed. If omitted, models that sample use a random seed unless their upstream default is fixed.

TTS streaming behavior

The current streaming TTS families do not all emit audio at the same granularity. VoxCPM2 and DotTTS can push audio from inside one generation request. Confucius4-TTS, NeuTTS, OmniVoice, and Supertonic 3 use pull events and emit a completed text segment/chunk at a time. The latter still reduces long-form time to first playback, but the first event waits for the first segment to finish.

Family	Delivery granularity	PCM output rate
voxcpm2	Intra-generation decoder chunks	48 kHz
dots_tts	Intra-generation streaming vocoder chunks	48 kHz
confucius4_tts	Completed text segments	22.05 kHz
neutts	Completed text segments	24 kHz
omnivoice	Completed model-planned text chunks	24 kHz
supertonic	Completed text chunks	44.1 kHz

The HTTP server exposes these as speech.audio.delta SSE events when the configured model uses mode: "streaming"; see the server API.

Qwen3 TTS

Qwen3 TTS supports reference voice cloning, voice design, and packaged custom voices. See Qwen3 models for the full Base, VoiceDesign, CustomVoice, ASR, and forced-alignment manual.

audiocpp_cli --task tts --family qwen3_tts --model models/Qwen3-TTS-12Hz-1.7B-Base --backend cuda --text "Hello from Qwen3 TTS." --voice-ref assets/resources/b.wav --out out.wav

Chatterbox

Chatterbox is a voice-clone TTS model with an audio-to-audio voice-conversion path. The upstream Chatterbox family also documents paralinguistic tag tokens in newer variants, but the current audio.cpp integration exposes voice cloning and voice conversion rather than a separate tag-control interface.

Field	Value
Family	chatterbox
Model directory	models/chatterbox
Tasks	clon, vc
Modes	offline
Languages	ar, da, de, el, en, es, fi, fr, hi, it, ko, ms, nl, no, pl, pt, sv, sw, tr
Voice input	Required reference WAV through --voice-ref; VC also requires source audio through --audio
Built-in voices	Not exposed by this integration

Voice clone:

audiocpp_cli --task clon --family chatterbox --model models/chatterbox --backend cuda --text "Hello from Chatterbox." --voice-ref assets/resources/b.wav --out out.wav

Voice conversion:

audiocpp_cli --task vc --family chatterbox --model models/chatterbox --backend cuda --audio assets/resources/a.wav --voice-ref assets/resources/b.wav --out converted.wav

Option	Values	Default	Meaning
--audio	WAV path	required for vc	Source speech for voice conversion.
--voice-ref	WAV path	required	Reference speaker audio for cloning, or target speaker audio for voice conversion.
--language	language code	en	Text language.
--text-chunk-size	integer chars	128	Long-form chunk size.
--guidance-scale	float	0.5	CFG strength.
--temperature	float	0.8	T3 sampling temperature.
--top-p	float	0.8	T3 nucleus sampling limit.
--repetition-penalty	float	2.0	T3 repetition penalty.
--max-tokens	integer	1000	Maximum generated T3 tokens per chunk.
--do-sample	true, false	true	Enable stochastic T3 sampling.

Confucius4-TTS

Confucius4-TTS is an experimental multilingual voice-cloning TTS model packaged as a standalone GGUF bundle. It supports offline generation and segment-level streaming audio output from a complete text request, using reference speech, language-aware text normalization, T2S semantic generation, S2A flow matching, style encoding, semantic audio features, and BigVGAN vocoding.

Field	Value
Family	confucius4_tts
GGUF model	models/Confucius4-TTS-GGUF/confucius4-tts-orig.gguf
Task	clon
Modes	offline, streaming
Languages	zh, en, ja, ko, de, fr, es, id, it, th, pt, ru, ms, vi
Voice input	Required reference WAV through --voice-ref
Built-in voices	Not exposed
Status	Experimental

Language support note: English (en) and Chinese (zh) are the currently validated and normalized paths. Other advertised language codes are experimental best-effort cross-language cloning paths; text frontend normalization is incomplete for them, so pronunciation and reading quality may vary.

Voice clone:

audiocpp_cli --task clon --family confucius4_tts --model models/Confucius4-TTS-GGUF/confucius4-tts-orig.gguf --backend cuda --language en --text "Hello from Confucius four TTS." --voice-ref assets/resources/b.wav --out out.wav

Streaming session:

audiocpp_cli --task clon --family confucius4_tts --model models/Confucius4-TTS-GGUF/confucius4-tts-orig.gguf --backend cuda --mode streaming --language en --text "Hello from the streaming path." --voice-ref assets/resources/b.wav --out out.wav

Option	Values	Default	Meaning
--voice-ref	WAV path	required	Reference speaker audio for cloning.
--language	language code	zh	Target synthesis language code.
--temperature	float	0.8	T2S sampling temperature.
--top-p	float	0.8	T2S nucleus sampling probability.
--top-k	integer	30	T2S top-k sampling limit.
--num-beams	integer	3	T2S beam count; use 1 for single-beam sampling.
--repetition-penalty	float	10.0	T2S repetition penalty.
--max-tokens	integer	1520	Maximum T2S semantic sequence length including prompt tokens.
--num-inference-steps	integer	25	S2A flow-matching step count.
--guidance-scale	float	0.7	S2A classifier-free guidance scale.
--text-chunk-size	integer tokens	80	Maximum text tokens per generated segment.
--text-chunk-mode	default, tag_aware, japanese, endline	default	Framework text chunking mode.
--request-option cross_fade_duration_sec=<seconds>	seconds	0.3	Cross-fade duration between generated segments.
--request-option edge_fade_duration_sec=<seconds>	seconds	0.1	Fade duration applied at segment edges.
--request-option edge_pad_duration_sec=<seconds>	seconds	0.1	Silence padding applied at segment edges.
--seed	integer	1234	Seed for T2S sampling and S2A noise initialization.
--session-option confucius4_tts.mem_saver=true|false	bool	false	Release staged graphs after request phases; default keeps them cached for reuse.

DramaBox

DramaBox is an experimental English expressive TTS and voice-cloning model packaged as a standalone GGUF bundle. It combines Gemma text conditioning, diffusion sampling, reference-audio conditioning, long-form chunking, and 48 kHz stereo output.

Field	Value
Family	dramabox
GGUF model	models/DramaBox-GGUF/dramabox-q8_0.gguf
Tasks	tts, clon
Modes	offline
Languages	en
Voice input	Optional reference WAV through --voice-ref
Built-in voices	Not exposed
Status	Experimental

Text-only speech:

audiocpp_cli --task tts --family dramabox --model models/DramaBox-GGUF/dramabox-q8_0.gguf --backend cuda --text "Hello from DramaBox." --out out.wav

Voice clone:

audiocpp_cli --task clon --family dramabox --model models/DramaBox-GGUF/dramabox-q8_0.gguf --backend cuda --text "Hello from DramaBox." --voice-ref assets/resources/b.wav --out out.wav

Older prebuilts that reject --task clon can use --task tts --voice-ref ...; the same reference-conditioning path is used.

Option	Values	Default	Meaning
--voice-ref / --target-voice	WAV path	not set	Reference voice for cloning; omitted requests text-only speech.
--request-option negative_prompt=<text>	string	built-in quality prompt	Negative text conditioning when classifier-free guidance is enabled.
--request-option duration_sec=<seconds>	seconds	0	Explicit target duration; 0 uses prompt-duration estimation.
--num-inference-steps	integer	30	Diffusion sampling steps.
--guidance-scale	float	2.5	Classifier-free guidance scale. Values greater than 1 enable CFG.
--request-option spatio_temporal_guidance_scale=<float>	float	1.5	Spatio-temporal guidance scale. Values greater than 0 enable STG.
--request-option duration_scale=<float>	float	1.1	Multiplier applied to the estimated prompt duration when duration_sec is 0.
--request-option reference_duration_sec=<seconds>	seconds	10.0	Reference voice crop/repeat duration.
--request-option guidance_rescale=auto|<number>	string	auto	Guidance rescale mode or explicit numeric value.
--request-option audio_chunk_threshold_sec=<seconds>	seconds	45.0	Estimated duration threshold that switches to long-form chunking.
--request-option audio_chunk_duration_sec=<seconds>	seconds	37.0	Target estimated duration for each long-form chunk.
--request-option cross_fade_duration_sec=<seconds>	seconds	0.05	Equal-power cross-fade between long-form chunks.
--seed	integer	42	Torch-compatible CUDA noise seed for diffusion sampling.
--session-option dramabox.perf_mode=off|flash_attention	enum	off	Attention implementation mode. off keeps the exact reference-query attention path; flash_attention enables the optimized path.
--session-option dramabox.mem_saver=true|false	bool	false	Release staged runtime graphs and weights immediately after each request phase to reduce peak and resident VRAM; default keeps components cached for reuse.

DotTTS

DotTTS is an experimental multilingual TTS and voice-cloning family with SOAR and MeanFlow GGUF packages. SOAR is the default download. See DotTTS for template, streaming, chunking, and full option details.

python3 tools/model_manager_v2.py install dots_tts_soar_q8_0

audiocpp_cli --task tts --family dots_tts \
  --model models/DotTTS-SOAR-GGUF/dots-tts-soar-q8_0.gguf \
  --backend cuda \
  --text "Our field team finished the morning inspection and prepared a concise update." \
  --voice-ref assets/resources/a.wav \
  --reference-text "This little work was finished in the year eighteen o three, and intended for immediate publication." \
  --request-option reference_duration_sec=5 \
  --out out.wav

For MeanFlow, install dots_tts_mf_q8_0 and use models/DotTTS-MF-GGUF/dots-tts-mf-q8_0.gguf with the same runtime options.

MioTTS

MioTTS is a 1.7B voice-clone TTS path that uses MioCodec for acoustic decoding. It requires a reference voice and a MioCodec model. Best-of-N candidate scoring can optionally use Qwen3-ASR.

Field	Value
Family	miotts
GGUF model	models/MioTTS-1.7B-GGUF/miotts-1.7b-q8_0.gguf
Required dependency	MioCodec through --session-option miotts.codec_model_path=<dir>
Task	tts
Modes	offline
Languages	Model auto-handles supported text languages; no explicit language selector is exposed
Voice input	Required reference WAV through --voice-ref
Built-in voices	Not exposed

audiocpp_cli --task tts --family miotts --model models/MioTTS-1.7B-GGUF/miotts-1.7b-q8_0.gguf --backend cuda --session-option miotts.codec_model_path=models/MioCodec-25Hz-44.1kHz-v2-GGUF/miocodec-25hz-44khz-v2-q8_0.gguf --text "Hello from MioTTS." --voice-ref assets/resources/b.wav --out out.wav

With best-of-N scoring, also provide a Qwen3-ASR model:

audiocpp_cli --task tts --family miotts --model models/MioTTS-1.7B-GGUF/miotts-1.7b-q8_0.gguf --backend cuda --session-option miotts.codec_model_path=models/MioCodec-25Hz-44.1kHz-v2-GGUF/miocodec-25hz-44khz-v2-q8_0.gguf --session-option miotts.best_of_n_asr_model_path=models/Qwen3-ASR-0.6B-GGUF/qwen3-asr-0.6b-q8_0.gguf --request-option miotts.best_of_n_enabled=true --request-option miotts.best_of_n=2 --text "Hello from MioTTS." --voice-ref assets/resources/b.wav --out out.wav

Option	Values	Default	Meaning
--voice-ref	WAV path	required	Reference speaker audio.
--text-chunk-size	integer chars	180	Long-form chunk size.
--max-tokens	integer	700	Maximum generated LM tokens per chunk.
--temperature	float	0.8	LM sampling temperature.
--top-k	integer	50	LM top-k sampling limit.
--top-p	float	1.0	LM nucleus sampling limit.
--repetition-penalty	float	1.0	LM repetition penalty.
--do-sample	true, false	true	Enable stochastic LM sampling.
--session-option miotts.codec_model_path=<dir>	directory	sibling MioCodec directory	MioCodec model used for acoustic decoding.
--request-option miotts.best_of_n_enabled=true|false	bool	false	Run best-of-N candidate selection.
--request-option miotts.best_of_n=<n>	integer	session default	Generate n candidates and select by ASR scoring.
--session-option miotts.best_of_n_default=<n>	integer	1	Default best-of-N candidate count.
--session-option miotts.best_of_n_max=<n>	integer	8	Maximum best-of-N candidate count.
--session-option miotts.best_of_n_language=auto|en|ja	enum	auto	Default language used when scoring candidates.
--session-option miotts.best_of_n_asr_model_path=<dir>	directory	sibling Qwen3-ASR directory	Qwen3-ASR model used for best-of-N scoring.

MOSS-TTS-Local

MOSS-TTS-Local is the larger local-transformer MOSS TTS path. It supports text-only speech and optional zero-shot voice cloning through the framework speaker-reference interface. See MOSS-TTS for tokenizer layout, sampling options, cache options, and Nano details.

Field	Value
Family	moss_tts_local
Model directory	models/MOSS-TTS-Local-Transformer-v1.5
Required codec layout	audio_tokenizer/ directory inside the model root
Task	tts, clon
Modes	offline
Languages	Model auto-handles supported languages; --language can pass a language hint
Voice input	Optional reference WAV through --voice-ref; transcript through --reference-text when known
Built-in voices	Not exposed

Text-only speech:

audiocpp_cli --task tts --family moss_tts_local --model /path/to/MOSS-TTS-Local-Transformer-v1.5 --backend cuda --text "Hello from MOSS-TTS-Local." --out out.wav

Voice clone:

audiocpp_cli --task clon --family moss_tts_local --model /path/to/MOSS-TTS-Local-Transformer-v1.5 --backend cuda --text "Hello from MOSS-TTS-Local." --voice-ref /path/to/reference.wav --reference-text "Reference transcript when available." --out out.wav

MOSS-TTS-Nano

MOSS-TTS-Nano is the smaller MOSS TTS path. It supports text-only continuation generation and voice cloning through the framework speaker-reference interface. See MOSS-TTS for tokenizer layout, sampling options, cache options, and Local details.

Field	Value
Family	moss_tts_nano
Model directory	models/MOSS-TTS-Nano-100M
Required codec layout	audio_tokenizer/ directory inside the model root
Task	tts, clon
Modes	offline
Languages	Model auto-handles supported languages
Voice input	Optional reference WAV through --voice-ref
Built-in voices	Not exposed

Text-only continuation:

audiocpp_cli --task tts --family moss_tts_nano --model /path/to/MOSS-TTS-Nano-100M --backend cuda --text "Hello from MOSS-TTS-Nano." --out out.wav

Voice clone:

audiocpp_cli --task clon --family moss_tts_nano --model /path/to/MOSS-TTS-Nano-100M --backend cuda --text "Hello from MOSS-TTS-Nano." --voice-ref /path/to/reference.wav --reference-text "Reference transcript when available." --out out.wav

NeuTTS

NeuTTS is an experimental English TTS family with built-in speaker prompts and emotion-token control. The default package is the standalone 2E GGUF. See NeuTTS for built-in voice ids, emotion options, streaming, and full option details.

python3 tools/model_manager_v2.py install neutts

audiocpp_cli --task tts --family neutts \
  --model models/NeuTTS-2E-GGUF/neutts-2e-orig.gguf \
  --backend cuda \
  --text "The release checklist is almost complete, and the baseline run looks healthy." \
  --request-option voice_id=emily \
  --request-option emotion=neutral \
  --out out.wav

OmniVoice

OmniVoice supports multilingual TTS, voice cloning, voice design, non-verbal tag tokens, long-form chunking, and chunked pseudo-streaming. See OmniVoice for the full guide.

Field	Value
Family	omnivoice
Model directory	models/OmniVoice
Task	tts
Modes	offline, streaming
Languages	600+ languages handled by the model
Voice input	--voice-ref plus optional --reference-text, or instruction text through --instruct
Built-in voices	Auto voice is supported by the model; CLI examples use clone or design for repeatability

Voice clone:

audiocpp_cli --task tts --family omnivoice --model models/OmniVoice --backend cuda --text "Hello from OmniVoice." --voice-ref assets/resources/b.wav --reference-text "Some call me nature. Others call me Mother Nature. I've been here for over 4.5 billion years. 22,500 times longer than you." --out out.wav

Voice design:

audiocpp_cli --task tts --family omnivoice --model models/OmniVoice --backend cuda --text "Hello from OmniVoice." --instruct "female, young adult, moderate pitch" --out out.wav

Streaming voice clone:

audiocpp_cli --task tts --mode streaming --family omnivoice --model models/OmniVoice --backend cuda --text "Hello from OmniVoice." --voice-ref assets/resources/b.wav --reference-text "Some call me nature. Others call me Mother Nature. I've been here for over 4.5 billion years. 22,500 times longer than you." --text-chunk-size 160 --out stream.wav --out-dir stream_chunks

OmniVoice streaming is pseudo streaming: audio.cpp emits audio chunk events from text chunks and returns a merged final WAV. Upstream Python does not expose model-native streaming. For server SSE examples, options, and tag controls, see OmniVoice.

PocketTTS

PocketTTS supports built-in voices and voice cloning. The upstream project also supports exported voice states for fast reuse; the CLI surface here exposes built-in voice ids and reference WAVs.

PocketTTS language selection is a model-load option. When the model path points at the PocketTTS root, the loader uses english unless you pass --load-option language=<name>. Kyutai's normal non-English PocketTTS releases are smaller distilled language models intended for the fast PocketTTS path. The _24l variants are larger 24-layer, undistilled preview models that can sound better but are slower. Kyutai currently publishes French only as french_24l, not as a normal distilled french language directory, so French is not listed as a normal PocketTTS language here.

Field	Value
Family	pocket_tts
Model directory	models/pocket-tts
Task	tts
Modes	offline
Languages	english, german, italian, portuguese, spanish
Voice input	Built-in voice id or reference WAV
Built-in voices	Voice ids depend on the downloaded language package; alba is used by the examples

Preset voice:

audiocpp_cli --task tts --family pocket_tts --model models/pocket-tts --backend cuda --text "Hello from PocketTTS." --voice-id alba --out out.wav

Voice clone:

audiocpp_cli --task tts --family pocket_tts --model models/pocket-tts --backend cuda --text "Hello from PocketTTS." --voice-ref assets/resources/b.wav --out out.wav

Option	Values	Default	Meaning
--load-option language=<name>	language package name	english	Select PocketTTS language package at load time.
--voice-id	packaged voice id	not set	Built-in voice id.
--voice-ref	WAV path	not set	Reference speaker audio for cloning.
--text-chunk-size	integer chars	256	Long-form chunk size.
--session-option pocket_tts.voice_state_cache_slots=<n>	integer slots	4	Prepared voice-state cache slots; set 0 to disable reuse.

VoxCPM2

VoxCPM2 supports plain TTS, voice design, controllable voice cloning, and an ultimate-clone style that uses both prompt audio and transcript. The CLI expresses voice design with the same text convention as the upstream examples: put the voice/style description in parentheses at the start of --text.

Field	Value
Family	voxcpm2
Model directory	models/VoxCPM2
Task	tts
Modes	offline, streaming
Languages	Model auto-handles supported languages
Voice input	Optional reference WAV; optional transcript through --reference-text
Built-in voices	Not exposed

Voice design:

audiocpp_cli --task tts --family voxcpm2 --model models/VoxCPM2 --backend cuda --text "(A young woman, gentle and clear voice)Hello from VoxCPM2." --out out.wav

Voice clone:

audiocpp_cli --task tts --family voxcpm2 --model models/VoxCPM2 --backend cuda --text "Hello from VoxCPM2." --voice-ref assets/resources/b.wav --out out.wav

Ultimate clone:

audiocpp_cli --task tts --family voxcpm2 --model models/VoxCPM2 --backend cuda --text "Hello from VoxCPM2." --voice-ref assets/resources/b.wav --reference-text "Some call me nature. Others call me Mother Nature. I've been here for over 4.5 billion years. 22,500 times longer than you." --out out.wav

Streaming output:

audiocpp_cli --task tts --family voxcpm2 --model models/VoxCPM2 --backend cuda --mode streaming --text "Hello from VoxCPM2." --out out.wav

Option	Values	Default	Meaning
--text "(style)content"	text	required	Voice design or style control.
--voice-ref	WAV path	not set	Reference speaker audio.
--reference-text	text	empty string	Transcript for ultimate-clone style prompting.
--mode	offline, streaming	offline	Full-output or streaming run mode.
--session-option voxcpm2.mem_saver=true|false	bool	false	Use tighter graph workspaces and release MiniCPM/AudioVAE request graphs after completion to reduce resident VRAM.
--session-option voxcpm2.prompt_cache_slots=<n>	integer	1	Prompt and prompt-audio embedding cache slots. Set to 0 to disable prompt caching.
--text-chunk-size	integer chars	2048	Long-form chunk size.
--text-chunk-mode	default, tag_aware, japanese, endline	tag_aware	Long-form chunking mode; keeps style/tag controls attached to chunks by default.
--max-tokens	integer	4096	Maximum generated AR tokens.
--num-inference-steps	integer	10	Flow matching steps.
--guidance-scale	float	2.0	CFG strength.

Higgs Audio v3 TTS

Higgs Audio v3 TTS is a voice-clone TTS model. The current integration uses the framework chunker for long text and keeps the reference prompt state in the model session.

Field	Value
Family	higgs_audio_tts
Model path	models/Higgs-Audio-v3-TTS-4B-GGUF/higgs-audio-v3-tts-4b-q8_0.gguf when installed through the model manager
Task	tts
Modes	offline
Languages	Model auto-handles supported languages
Voice input	Reference WAV through --voice-ref; transcript through --reference-text when known
Built-in voices	Not exposed

audiocpp_cli --task tts --family higgs_audio_tts --model models/Higgs-Audio-v3-TTS-4B-GGUF/higgs-audio-v3-tts-4b-q8_0.gguf --backend cuda --text "Hello from Higgs Audio." --voice-ref assets/resources/b.wav --reference-text "Some call me nature. Others call me Mother Nature. I've been here for over 4.5 billion years. 22,500 times longer than you." --out out.wav

The model manager installs the Q8_0 standalone GGUF package by default:

python3 tools/model_manager_v2.py install --models-root models higgs_audio_tts_4b_q8_0

Option	Values	Default	Meaning
--voice-ref	WAV path	required	Reference speaker audio.
--reference-text	text	empty string	Transcript for reference audio.
--text-chunk-size	integer chars	1024	Long-form chunk size.
--max-tokens	integer	2048	Maximum generated AR tokens per chunk.
--temperature	float	0.8	AR sampling temperature.
--top-k	integer	30	AR top-k sampling limit. The narrower default is less prone to premature EOC than the Python client's 50.
--top-p	float	0.8	AR nucleus sampling limit. The Python client's unfiltered equivalent is 1.0.
--repetition-penalty	float	1.1	Accepted for Python API compatibility; Higgs audio-code sampling does not consume it.

Fish Audio S2 Pro

Fish Audio S2 Pro is a TTS and reference voice-clone model. The integration uses the framework text chunker for long-form input, caches prepared reference audio in the session, and supports GGUF loading through the package spec path.

Field	Value
Family	fish_audio
Model path	models/Fish-Audio-S2-Pro-GGUF/fish-audio-s2-pro-q8_0.gguf when installed through the model manager
Task	tts
Modes	offline
Languages	Model auto-handles language; tested paths cover English and Chinese-style prompts
Voice input	Optional reference WAV through --voice-ref; transcript through --reference-text when known
Built-in voices	Not exposed

Text-to-speech:

audiocpp_cli --task tts --family fish_audio --model models/Fish-Audio-S2-Pro-GGUF/fish-audio-s2-pro-q8_0.gguf --backend cuda --text "Hello from Fish Audio." --out out.wav

Reference voice clone:

audiocpp_cli --task tts --family fish_audio --model models/Fish-Audio-S2-Pro-GGUF/fish-audio-s2-pro-q8_0.gguf --backend cuda --text "The final render is ready for review." --voice-ref assets/resources/b.wav --reference-text "Some call me nature. Others call me Mother Nature. I've been here for over 4.5 billion years. 22,500 times longer than you." --out out.wav

The model manager installs the Q8_0 standalone GGUF package by default:

python3 tools/model_manager_v2.py install --models-root models fish_audio_s2_pro_q8_0

Option	Values	Default	Meaning
--voice-ref	WAV path	not set	Reference speaker audio for voice cloning.
--reference-text	text	empty string	Transcript for reference audio.
--max-new-tokens	integer	1024	Maximum generated semantic tokens per chunk. 0 uses the default.
--text-chunk-size	integer chars	200	Long-form chunk size.
--text-chunk-mode	default, tag_aware, japanese, endline	default	Framework text chunking mode.
--temperature	float	0.8	Sampling temperature.
--top-k	integer	30	Top-k sampling limit.
--top-p	float	0.8	Nucleus sampling limit.
--seed	integer	random when omitted	Sampling seed for reproducible output.
--session-option fish_audio.mem_saver=true|false	bool	false	Release cached AR runtime graphs after each request.
--session-option fish_audio.reference_cache_slots=<n>	integer	1	Prepared reference-audio cache slots.
--session-option fish_audio.weight_type=<type>	native, f32, f16, bf16, q8_0	native	AR matmul weight storage type.
--session-option fish_audio.codec_weight_type=<type>	native, f32, f16, q8_0	native	Codec conv/matmul weight storage type.

IndexTTS2

IndexTTS2 is a Chinese and English TTS model with voice cloning and expressive emotion controls. It requires a speaker reference through the framework --voice-ref path.

Field	Value
Family	index_tts2
Model directory	models/IndexTTS-2
Task	tts, clon
Modes	offline
Languages	zh, en
Voice input	Required reference WAV through --voice-ref
Built-in voices	Not exposed

Voice clone:

audiocpp_cli --task clon --family index_tts2 --model /path/to/IndexTTS-2 --backend cuda --language en --text "Hello from IndexTTS2." --voice-ref /path/to/reference.wav --out out.wav

Emotion text:

audiocpp_cli --task tts --family index_tts2 --model /path/to/IndexTTS-2 --backend cuda --language zh --text "今天的演示会更有情绪。" --voice-ref /path/to/reference.wav --emotion "你吓死我了！你是鬼吗？" --request-option emotion_alpha=0.6 --out out.wav

Option	Values	Default	Meaning
--voice-ref	WAV path	required	Reference speaker audio.
--language	zh, en	empty	Text language label.
--emotion	text	not set	Emotion-text conditioning through the framework style field.
--request-option emotion_alpha=<float>	float in [0, 1]	1.0	Blend strength for explicit emotion conditioning.
--request-option emotion_vector=<v0,...,v7>	8 floats	not set	Explicit emotion vector.
--request-option use_emotion_text=true|false	bool	false	Infer emotion from text.
--request-option use_random_emotion=true|false	bool	false	Use random emotion weights in the emotion mixer.
--request-option interval_silence_ms=<n>	milliseconds	200	Silence inserted between generated text chunks.
--text-chunk-size	characters	not set	Optional framework outer text chunk size. When omitted, IndexTTS2 keeps its internal tokenizer segmentation.
--text-chunk-mode	default, tag_aware, japanese, endline	default	Framework chunking mode used only when --text-chunk-size is set.
--max-tokens	integer	model default	Maximum generated GPT mel tokens.
--temperature	float	model default	GPT sampling temperature.
--top-p	float	model default	GPT nucleus sampling limit.
--top-k	integer	model default	GPT top-k sampling limit.
--repetition-penalty	float	model default	GPT repetition penalty.
--do-sample	true, false	model default	Enable stochastic GPT sampling.
--session-option index_tts2.mem_saver=true|false	bool	false	Release staged reference and conditioning graphs after request phases.
--session-option index_tts2.weight_type=native|f32|f16|bf16|q8_0	enum	native	Matmul weight storage type.
--session-option index_tts2.conv_weight_type=native|f32|f16	enum	native	Convolution weight storage type.
--session-option index_tts2.speaker_cache_slots=<n>	integer slots	1	Prepared speaker-reference cache slots; set 0 to disable reuse.
--session-option index_tts2.emotion_cache_slots=<n>	integer slots	1	Prepared emotion-reference cache slots; set 0 to disable reuse.
--session-option index_tts2.emotion_text_cache_slots=<n>	integer slots	1	Emotion-text weight cache slots; set 0 to disable reuse.
--session-option index_tts2.gpt_graph_arena_mb=<n>	MB	model default	GPT graph arena size.
--session-option index_tts2.s2mel_graph_arena_mb=<n>	MB	model default	S2Mel graph arena size.
--session-option index_tts2.reference_graph_arena_mb=<n>	MB	model default	Reference encoder and codec graph arena size.
--session-option index_tts2.emotion_text_prefill_graph_arena_mb=<n>	MB	model default	Emotion-text prefill graph arena size.
--session-option index_tts2.emotion_text_decode_graph_arena_mb=<n>	MB	model default	Emotion-text cached-step graph arena size.
--session-option index_tts2.emotion_text_max_tokens=<n>	tokens	256	Maximum generated tokens for emotion-text classification; old name index_tts2.emotion_text_max_new_tokens is still accepted.
--session-option index_tts2.weight_context_mb=<n>	MB	32	Shared ggml weight metadata context size.

IndexTTS2.5

IndexTTS2.5 is IndexTeam/bilibili's multilingual zero-shot TTS model (released 2026-07): a 0.8B GPT (autoregressive) + DiT CFM + BigVGAN stack that keeps IndexTTS2's timbre-emotion decoupling and adds Japanese, Spanish, and Arabic on top of Chinese and English. It requires a speaker reference through the framework --voice-ref path. Inline <文字|发音> pronunciation overrides (pinyin, CMU phonemes, or kana) are supported. Upstream weights live at IndexTeam/IndexTTS-2.5; the reference implementation is index-tts/index-tts branch indextts-2.5.

IndexTTS2.5 is implemented as a variant of the index_tts2 family rather than a separate family: both variants share the audio features, wav2vec2bert, Qwen emotion, style encoder, BigVGAN vocoder, S2Mel, and the GPT decode/cache code, while the tokenizer (SentencePiece vs multilingual tiktoken), GPT speaker conditioning (conditioning encoder + perceiver vs CAMPPlus spk_emb_proj + lang_embedding), and the semantic-codec decode path (v2.5 adds a 2x nearest upsample + up conv) are selected per variant from the model config version field ("2.5"). All IndexTTS2 session options (index_tts2.*) apply to both variants.

Field	Value
Family	index_tts2 (the 2.5 variant is selected from the model config version field; no separate family)
Model directory	models/IndexTTS2.5-GGUF (default GGUF package index_tts2_5_q8_0; index_tts2_5_f16 and index_tts2_5_orig also available)
Task	tts, clon
Modes	offline
Languages	zh, en, ja, es, ar
Voice input	Required reference WAV through --voice-ref
Built-in voices	Not exposed

Voice clone:

audiocpp_cli --task clon --family index_tts2 --model /path/to/IndexTTS2.5-GGUF --backend cuda --text "Hello from IndexTTS2.5." --voice-ref /path/to/reference.wav --out out.wav

Emotion text:

audiocpp_cli --task tts --family index_tts2 --model /path/to/IndexTTS2.5-GGUF --backend cuda --text "今天的演示会更有情绪。" --voice-ref /path/to/reference.wav --emotion "你吓死我了！你是鬼吗？" --request-option emotion_alpha=0.6 --out out.wav

The language request option selects the text language (auto, zh, en, ja, es, ar, or any tokenizer language code). The default auto picks zh when the text contains Han characters and en otherwise, so mixed Japanese/Spanish/Arabic text should set --request-option language=ja|es|ar explicitly.

Emotion conditioning supports all three IndexTTS2 paths: an emotion reference WAV through --audio, an explicit emotion_vector, and Qwen-based emotion-text classification through --emotion / use_emotion_text. Known limitation: the NeMo text normalizers for Japanese and Spanish are not ported, so ja/es input text is passed through without upstream-style normalization.

License: IndexTTS-2.5 weights are distributed under the bilibili Model Use License, which is not OSI-approved. It requires separate commercial authorization when monthly active users exceed 100 million or annual revenue exceeds 1 billion RMB, and it forbids using model outputs to improve other AI models. Check the upstream repository for the full terms before redistribution or commercial use.

Option	Values	Default	Meaning
--voice-ref	WAV path	required	Reference speaker audio.
--request-option language=<code>	auto, zh, en, ja, es, ar, ...	auto	Text language hint; auto infers zh when the text contains Han characters, otherwise en.
--emotion	text	not set	Emotion-text conditioning through the framework style field.
--request-option emotion_alpha=<float>	float in [0, 1]	1.0	Blend strength for explicit emotion conditioning.
--request-option emotion_vector=<v0,...,v7>	8 floats	not set	Explicit emotion vector.
`--request-option use_emotion_text=true	false`	bool	false
`--request-option use_random_emotion=true	false`	bool	false
--request-option interval_silence_ms=<n>	milliseconds	200	Silence inserted between generated text chunks.
--text-chunk-size	characters	not set	Optional framework outer text chunk size. When omitted, IndexTTS2.5 keeps its internal tokenizer segmentation.
--text-chunk-mode	default, tag_aware, japanese, endline	default	Framework chunking mode used only when --text-chunk-size is set.
--max-tokens	integer	1500	Maximum generated GPT mel tokens.
--temperature	float	0.8	GPT sampling temperature.
--top-p	float	0.8	GPT nucleus sampling limit.
--top-k	integer	30	GPT top-k sampling limit.
--repetition-penalty	float	10.0	GPT repetition penalty.
--do-sample	true, false	true	Enable stochastic GPT sampling.
--request-option length_penalty=<float>	float	0.0	GPT beam-search length penalty.
--request-option num_beams=<n>	integer	3	GPT beam count.
`--session-option index_tts2.mem_saver=true	false`	bool	false
`--session-option index_tts2.weight_type=native	f32	f16	bf16
`--session-option index_tts2.conv_weight_type=native	f32	f16`	enum
--session-option index_tts2.speaker_cache_slots=<n>	integer slots	1	Prepared speaker-reference cache slots; set 0 to disable reuse.
--session-option index_tts2.emotion_cache_slots=<n>	integer slots	1	Prepared emotion-reference cache slots; set 0 to disable reuse.
--session-option index_tts2.emotion_text_cache_slots=<n>	integer slots	1	Emotion-text weight cache slots; set 0 to disable reuse.
--session-option index_tts2.gpt_graph_arena_mb=<n>	MB	model default	GPT graph arena size.
--session-option index_tts2.s2mel_graph_arena_mb=<n>	MB	model default	S2Mel graph arena size.
--session-option index_tts2.reference_graph_arena_mb=<n>	MB	model default	Reference encoder and codec graph arena size.
--session-option index_tts2.emotion_text_prefill_graph_arena_mb=<n>	MB	model default	Emotion-text prefill graph arena size.
--session-option index_tts2.emotion_text_decode_graph_arena_mb=<n>	MB	model default	Emotion-text cached-step graph arena size.
--session-option index_tts2.emotion_text_max_tokens=<n>	tokens	256	Maximum generated tokens for emotion-text classification; old name index_tts2.emotion_text_max_new_tokens is still accepted.
--session-option index_tts2.weight_context_mb=<n>	MB	32	Shared ggml weight metadata context size.

Converting From Upstream Weights

tools/convert_index_tts2_5.py turns an official IndexTeam/IndexTTS-2.5 snapshot (the .pth checkpoints) into the Safetensors staging layout the engine expects, and prints (or runs) the matching audiocpp_gguf command. The w2v-bert-2.0, CAMPPlus, and BigVGAN checkpoints are auto-detected under <model-dir>/hf_cache/ (run the official inference once to populate it) and each has an explicit override flag:

python tools/convert_index_tts2_5.py \
    --model-dir /path/to/IndexTTS-2.5 \
    --output-dir /path/to/staging \
    --run-converter /path/to/audiocpp_gguf --type q8_0

Pass --native-dir /path/to/IndexTTS-2.5-native to also emit a directly loadable native Safetensors model directory (hardlinked from the staging files), no GGUF conversion required.

The script repackages the checkpoints the loader needs (unwraps the s2mel.pth/codec.pth container keys, prefixes CAMPPlus tensors with speaker_encoder., strips BigVGAN's generator. prefix, wraps the feat1/feat2.pt matrices as a single tensor) and assembles the sidecar root/ (config, tiktoken vocabulary, auxiliary model configs) that gets embedded into the GGUF. The staged config.yaml has its version field normalized to "2.5" (the official snapshot ships version: 2.0); the engine uses that field to select the IndexTTS2 family variant.

Irodori-TTS

Irodori-TTS is Japanese TTS under --family irodori_tts. v4 Small is the preferred GGUF-first package and supports no-reference speech, reference-conditioned speech, and instruction-based voice design in one checkpoint. The older 500M v3 and 600M v3 VoiceDesign packages remain supported for existing users. See Irodori-TTS for v3/v4 differences, GGUF variants, options, and compatibility aliases.

OuteTTS

OuteTTS 1.0 1B is a community model for 24 kHz TTS and voice cloning. The model manager installs the standalone Q8 GGUF package by default:

python tools/model_manager_v2.py install outetts_1_0_1b_q8_0 --models-root models

Quick start:

audiocpp_cli --task tts --family outetts \
  --model models/Llama-OuteTTS-1.0-1B_Q8/Llama-OuteTTS-1.0-1B_Q8.gguf \
  --backend cuda --text "Hello from OuteTTS." \
  --max-tokens 1024 --out out.wav

Voice clone quick start:

audiocpp_cli --task clon --family outetts \
  --model models/Llama-OuteTTS-1.0-1B_Q8/Llama-OuteTTS-1.0-1B_Q8.gguf \
  --backend cuda \
  --voice-ref reference.wav \
  --reference-text "The exact words spoken in reference.wav." \
  --request-option reference_language=en \
  --text "This sentence uses the cloned voice." \
  --max-tokens 1024 --out cloned.wav

See OuteTTS community model usage for cloning notes, GGUF packing, all options, and validation details.

GLM-TTS

GLM-TTS is a community zero-shot Chinese and English speech-synthesis model. Both the tts and clon routes require a clean reference WAV and its exact transcript:

python tools/model_manager_v2.py install glm_tts --models-root models

audiocpp_cli --task clon --family glm_tts \
  --model models/GLM-TTS-Q8/GLM-TTS_Q8.gguf --backend cuda \
  --voice-ref reference.wav \
  --reference-text "The exact words spoken in reference.wav." \
  --text "Hello from GLM-TTS." \
  --seed 0 --out glm_tts.wav

See the GLM-TTS community model guide for standalone GGUF packaging, controls, and validation results.

Inflect v2

Inflect Micro v2 is a compact English offline TTS model with a native GGML runtime. The model manager defaults to the standalone GGUF package; the original source/conversion path remains documented in the community guide. Inflect requires an external eSpeak-ng installation:

python3 tools/model_manager_v2.py install inflect_micro_v2_orig --models-root models

audiocpp_cli --task tts --family inflect_v2 \
  --model models/Inflect-Micro-v2-GGUF/inflect-micro-v2-orig.gguf --backend cuda \
  --text "Hello from Inflect Micro version two." \
  --request-option speaking_rate=1.0 \
  --request-option variation=0.667 \
  --seed 0 --out inflect.wav

See the Inflect v2 community model guide for eSpeak-ng paths, long-form behavior, source/conversion instructions, and limitations.

Supertonic

Supertonic 3 is a preset-voice multilingual TTS model. It does not use external speaker references in the current integration.

Field	Value
Family	supertonic
Model directory	models/supertonic-3
Task	tts
Modes	offline, streaming
Languages	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
Voice input	Built-in preset voice id
Built-in voices	M1-M5, F1-F5

audiocpp_cli --task tts --family supertonic --model /path/to/supertonic-3 --backend cuda --language en --text "Hello from Supertonic." --voice-id M1 --out out.wav

Streaming output:

audiocpp_cli --task tts --family supertonic --model /path/to/supertonic-3 --backend cuda --mode streaming --language en --text "Hello from Supertonic." --voice-id M1 --out out.wav

Option	Values	Default	Meaning
--voice-id	M1-M5, F1-F5	M1	Preset voice.
--language	language code	en	Text language.
--num-inference-steps	integer	8	Flow denoising steps.
--request-option speaking_rate=<float>	float	1.05	Speech speed multiplier.
--seed	integer	1234	Noise seed.
--text-chunk-size	characters	300, or 120 for ko/ja	Framework long-form text chunk size.
--text-chunk-mode	default, tag_aware, japanese, endline	default	Framework long-form text chunking mode.
--session-option supertonic.weight_type=native|f32|f16|bf16|q8_0	enum	native	Weight storage type.
--session-option supertonic.style_cache_slots=<n>	integer slots	4	Preset voice style cache slots; set 0 to disable reuse.

VibeVoice

VibeVoice is a long-form multi-speaker TTS model, available in 1.5B and 7B sizes. Prompts use speaker-labeled lines, and speaker reference WAVs are provided in the same order as the speaker ids.

Field	Value
Family	vibevoice
Model directory	models/VibeVoice-1.5B (or models/VibeVoice-7B)
Task	tts
Modes	offline
Languages	Model auto-handles supported languages
Voice input	Up to four speaker reference WAVs through voice_samples
Text format	Lines like Speaker 1: ... Speaker 2: ...; ids are normalized internally
Long-form	No text chunking; generation uses the model long-form path
LoRA	Optional PEFT decoder adapter through --load-option vibevoice.lora

Both sizes share the same CLI surface and the same Qwen2.5 tokenizer; the 7B is simply larger (hidden size 3584 vs 1536) and needs a matching 7B LoRA if one is used.

audiocpp_cli --task tts --family vibevoice --model models/VibeVoice-1.5B --backend cuda --text "Speaker 1: Hello. Speaker 2: Nice to meet you." --request-option voice_samples=assets/resources/a.wav,assets/resources/b.wav --out out.wav

Option	Values	Default	Meaning
--request-option voice_samples=a.wav,b.wav	comma-separated WAVs	not set	Speaker reference WAVs, ordered by speaker id.
--guidance-scale	float	1.3	Classifier-free guidance scale.
--num-inference-steps	integer	10	Diffusion steps per audio chunk.
--max-tokens	integer, 0 for unlimited	0	Maximum generated decoder tokens.
--request-option max_length_times=<float>	float	2.0	Generation length multiplier.
--do-sample	true, false	false	Enable stochastic decoder sampling.
--temperature	float	1.0	Decoder sampling temperature.
--top-k	integer	50	Decoder top-k sampling limit.
--top-p	float	1.0	Decoder nucleus sampling limit.
--load-option vibevoice.lora=<path>	fine-tune adapter dir	not set	Overlay a fine-tune at load time: the language-model LoRA is delta-merged into the decoder linears, and the diffusion head and acoustic/semantic connectors (when present in the adapter dir) replace their base tensors. Dims must match the base model size.
--load-option vibevoice.lora_scale=<float>	float	lora_alpha / r	Override the LoRA merge scale from adapter_config.json.

The adapter follows the PEFT training layout: adapter_model.safetensors + adapter_config.json for the language-model LoRA, plus optional diffusion_head/model.safetensors (or diffusion_head_full.bin), acoustic_connector/pytorch_model.bin, and semantic_connector/pytorch_model.bin for the fully fine-tuned components. Everything is applied at load time, so it composes with the vibevoice.*_weight_type quantization options and adds no per-step cost; the overlay is logged with --log. Use a 1.5B adapter with VibeVoice-1.5B and a 7B adapter with VibeVoice-7B; a size mismatch is rejected with a descriptive error. The same option may instead be passed as --session-option vibevoice.lora (but not via both at once).

For backend weight-type controls, use audiocpp_cli --inspect --model <model-dir> --family <family>.