TTS Browser Demos

sherpa-onnx TTS running entirely in your browser via WebAssembly

Kitten

Kitten Nano EN v0.2

Fast, lightweight English TTS. fp16 model (~40 MB). Great for quick testing.

Fast · English

Kitten Nano EN v0.1

Older Nano version. fp16 model (~40 MB). For comparison with v0.2.

Fast · English

Kitten Mini EN

Larger/better quality than Nano. fp16 model (~60 MB). 8 speakers.

English · 8 speakers

Kokoro

Kokoro v1.0 Multi-Lang (int8)

9 languages, 53 speakers. int8 model (~126 MB). Includes Japanese, Chinese, Korean, and more.

Multi-Lang · 53 speakers

Bug: silent audio (NaN) — upstream issue #2236

Kokoro v1.1 Multi-Lang (int8)

Newer version, 9 languages, 103 speakers. int8 model (~130 MB).

Multi-Lang · 103 speakers

Bug: may have same NaN issue as v1.0 int8

Kokoro v1.0 Multi-Lang (fp32)

Non-quantized v1.0, 9 languages, 53 speakers. fp32 model (~310 MB).

Multi-Lang · 53 speakers

fp32 — may work around int8 NaN bug

Kokoro v1.1 Multi-Lang (fp32)

Non-quantized v1.1, 9 languages, 103 speakers. fp32 model (~310 MB).

Multi-Lang · 103 speakers

fp32 — may work around int8 NaN bug

Kokoro EN v0.19 (fp32)

English-only Kokoro. fp32 model (~330 MB). 11 speakers.

English · 11 speakers

Kokoro EN v0.19 (int8)

English-only Kokoro. int8 model (~90 MB). 11 speakers.

English · 11 speakers

May have NaN issue (int8)

Matcha

Matcha EN (LJSpeech)

English Matcha TTS. 22kHz with vocos vocoder. ~125 MB.

English · 22kHz

Matcha ZH (Baker)

Chinese Matcha TTS. 22kHz with vocos vocoder. ~125 MB.

Chinese · 22kHz

Matcha ZH-EN

Chinese-English bilingual Matcha TTS. 16kHz with vocos vocoder. ~130 MB.

ZH-EN · 16kHz

Matcha FA-EN (Khadijah)

Farsi-English Matcha TTS. 22kHz, female voice. ~125 MB.

FA-EN · 22kHz

Matcha FA-EN (Musa)

Farsi-English Matcha TTS. 22kHz, male voice. ~125 MB.

FA-EN · 22kHz

ZipVoice

ZipVoice Distill ZH-EN (int8)

Distilled ZipVoice, Chinese-English. int8 model + vocos vocoder. ~160 MB.

ZH-EN · 24kHz

Not implemented: ZipVoice generate() not yet supported in WASM build

ZipVoice Distill ZH-EN (fp32)

Distilled ZipVoice, Chinese-English. fp32 model + vocos vocoder. ~510 MB.

ZH-EN · 24kHz

Not implemented: ZipVoice generate() not yet supported in WASM build

Pocket

Pocket TTS (int8)

Compact Pocket TTS. int8 model (~94 MB). Novel 5-component architecture.

Compact · int8

Not implemented: Pocket TTS generate() not yet supported in WASM build

Pocket TTS (fp32)

Full-precision Pocket TTS. fp32 model (~160 MB). Novel 5-component architecture.

Full precision · fp32

Not implemented: Pocket TTS generate() not yet supported in WASM build

Piper

Piper EN (US)

English multi-speaker VITS. 904 speakers. fp32 / int8 precision selectable.

English · 904 speakers

Piper ZH (CN)

Chinese (Mandarin) VITS. Single speaker. fp32 / int8 precision selectable.

Chinese · 1 speaker

Piper DE

German emotional VITS. Single speaker. fp32 / int8 precision selectable.

German · 1 speaker

Piper AR (Arabic)

AR (Arabic) VITS. 1 speaker. fp32 / int8 selectable.

AR (Arabic) · 1 speaker

Piper CA (Catalan)

CA (Catalan) VITS. 1 speaker. fp32 / int8 selectable.

CA (Catalan) · 1 speaker

Piper CS (Czech)

CS (Czech) VITS. 1 speaker. fp32 / int8 selectable.

CS (Czech) · 1 speaker

Piper CY (Welsh)

CY (Welsh) VITS. 1 speaker. fp32 / int8 selectable.

CY (Welsh) · 1 speaker

Piper DA (Danish)

DA (Danish) VITS. 1 speaker. fp32 / int8 selectable.

DA (Danish) · 1 speaker

Piper EL (Greek)

EL (Greek) VITS. 1 speaker. fp32 / int8 selectable.

EL (Greek) · 1 speaker

Piper EN (GB)

EN (GB) VITS. 1 speaker. fp32 / int8 selectable.

EN (GB) · 1 speaker

Piper ES (Spain)

ES (Spain) VITS. 1 speaker. fp32 / int8 selectable.

ES (Spain) · 1 speaker

Piper ES (Argentina)

ES (Argentina) VITS. 1 speaker. fp32 / int8 selectable.

ES (Argentina) · 1 speaker

Piper ES (Mexico)

ES (Mexico) VITS. 1 speaker. fp32 / int8 selectable.

ES (Mexico) · 1 speaker

Piper FA (Persian)

FA (Persian) VITS. 1 speaker. fp32 / int8 selectable.

FA (Persian) · 1 speaker

Piper FA-EN (Bilingual)

FA-EN (Bilingual) VITS. 1 speaker. fp32 / int8 selectable.

FA-EN (Bilingual) · 1 speaker

Piper FI (Finnish)

FI (Finnish) VITS. 1 speaker. fp32 / int8 selectable.

FI (Finnish) · 1 speaker

Piper FR (French)

FR (French) VITS. 1 speaker. fp32 / int8 selectable.

FR (French) · 1 speaker

Piper HI (Hindi)

HI (Hindi) VITS. 1 speaker. fp32 / int8 selectable.

HI (Hindi) · 1 speaker

Piper HU (Hungarian)

HU (Hungarian) VITS. 1 speaker. fp32 / int8 selectable.

HU (Hungarian) · 1 speaker

Piper ID (Indonesian)

ID (Indonesian) VITS. 1 speaker. fp32 / int8 selectable.

ID (Indonesian) · 1 speaker

Piper IS (Icelandic)

IS (Icelandic) VITS. 1 speaker. fp32 / int8 selectable.

IS (Icelandic) · 1 speaker

Piper IT (Italian)

IT (Italian) VITS. 1 speaker. fp32 / int8 selectable.

IT (Italian) · 1 speaker

Piper KA (Georgian)

KA (Georgian) VITS. 1 speaker. fp32 / int8 selectable.

KA (Georgian) · 1 speaker

Piper KK (Kazakh)

KK (Kazakh) VITS. 2 speakers. fp32 / int8 selectable.

KK (Kazakh) · 2 speakers

Piper LB (Luxembourgish)

LB (Luxembourgish) VITS. 1 speaker. fp32 / int8 selectable.

LB (Luxembourgish) · 1 speaker

Piper LV (Latvian)

LV (Latvian) VITS. 1 speaker. fp32 / int8 selectable.

LV (Latvian) · 1 speaker

Piper ML (Malayalam)

ML (Malayalam) VITS. 1 speaker. fp32 / int8 selectable.

ML (Malayalam) · 1 speaker

Piper NE (Nepali)

NE (Nepali) VITS. 1 speaker. fp32 / int8 selectable.

NE (Nepali) · 1 speaker

Piper NL (Netherlands)

NL (Netherlands) VITS. 1 speaker. fp32 / int8 selectable.

NL (Netherlands) · 1 speaker

Piper NL (Belgium)

NL (Belgium) VITS. 1 speaker. fp32 / int8 selectable.

NL (Belgium) · 1 speaker

Piper NO (Norwegian)

NO (Norwegian) VITS. 1 speaker. fp32 / int8 selectable.

NO (Norwegian) · 1 speaker

Piper PL (Polish)

PL (Polish) VITS. 1 speaker. fp32 / int8 selectable.

PL (Polish) · 1 speaker

Piper PT (Portugal)

PT (Portugal) VITS. 1 speaker. fp32 / int8 selectable.

PT (Portugal) · 1 speaker

Piper PT (Brazil)

PT (Brazil) VITS. 1 speaker. fp32 / int8 selectable.

PT (Brazil) · 1 speaker

Piper RO (Romanian)

RO (Romanian) VITS. 1 speaker. fp32 / int8 selectable.

RO (Romanian) · 1 speaker

Piper RU (Russian)

RU (Russian) VITS. 1 speaker. fp32 / int8 selectable.

RU (Russian) · 1 speaker

Piper SK (Slovak)

SK (Slovak) VITS. 1 speaker. fp32 / int8 selectable.

SK (Slovak) · 1 speaker

Piper SL (Slovenian)

SL (Slovenian) VITS. 1 speaker. fp32 / int8 selectable.

SL (Slovenian) · 1 speaker

Piper SR (Serbian)

SR (Serbian) VITS. 1 speaker. fp32 / int8 selectable.

SR (Serbian) · 1 speaker

Piper SV (Swedish)

SV (Swedish) VITS. 1 speaker. fp32 / int8 selectable.

SV (Swedish) · 1 speaker

Piper SW (Swahili)

SW (Swahili) VITS. 1 speaker. fp32 / int8 selectable.

SW (Swahili) · 1 speaker

Piper TR (Turkish)

TR (Turkish) VITS. 1 speaker. fp32 / int8 selectable.

TR (Turkish) · 1 speaker

Piper UK (Ukrainian)

UK (Ukrainian) VITS. 1 speaker. fp32 / int8 selectable.

UK (Ukrainian) · 1 speaker

Piper VI (Vietnamese)

VI (Vietnamese) VITS. 1 speaker. fp32 / int8 selectable.

VI (Vietnamese) · 1 speaker

Piper NL-BE RDH (X-Low)

Dutch VITS. 1 speaker. ~25 MB per variant. fp32 only.

Dutch · 1 speaker

Piper CA Upc Ona (X-Low)

Catalan VITS. 1 speaker. ~25 MB per variant.

Catalan · 1 speaker

Piper CA Upc Pau (X-Low)

Catalan VITS. 1 speaker. ~25 MB per variant.

Catalan · 1 speaker

Piper DE Eva K (X-Low)

German VITS. 1 speaker. ~25 MB per variant.

German · 1 speaker

Piper ES Carlfm (X-Low)

Spanish VITS. 1 speaker. ~25 MB per variant.

Spanish · 1 speaker

Piper IT Riccardo (X-Low)

Italian VITS. 1 speaker. ~25 MB per variant.

Italian · 1 speaker

Piper KK Iseke (X-Low)

Kazakh VITS. 1 speaker. ~25 MB per variant.

Kazakh · 1 speaker

Piper KK Raya (X-Low)

Kazakh VITS. 1 speaker. ~25 MB per variant.

Kazakh · 1 speaker

Piper NL-BE Nathalie (X-Low)

Dutch VITS. 1 speaker. ~25 MB per variant.

Dutch · 1 speaker

Piper UK Lada (X-Low)

Ukrainian VITS. 1 speaker. ~25 MB per variant.

Ukrainian · 1 speaker

Piper NE Google (X-Low)

Nepali multi-speaker VITS. ~32 MB per variant.

Nepali · multi-speaker

Piper VI Vivos (X-Low)

Vietnamese multi-speaker VITS. ~32 MB per variant.

Vietnamese · multi-speaker

Piper FR Siwis (Low)

French VITS. 1 speaker. ~25 MB per variant.

French · 1 speaker

Piper AR Kareem (Low)

Arabic VITS. 1 speaker. ~64 MB per variant.

Arabic · 1 speaker

Piper CS Jirka (Low)

Czech VITS. 1 speaker. ~64 MB per variant.

Czech · 1 speaker

Piper DE Glados (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Glados Turret (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Karlsson (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Kerstin (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Pavoque (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Ramona (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Thorsten (Low)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper EN-GB Alan (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN-GB South Female (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Amy (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Danny (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Kathleen (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Lessac (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Ryan (Low)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper FI Harri (Low)

Finnish VITS. 1 speaker. ~64 MB per variant.

Finnish · 1 speaker

Piper FR Gilles (Low)

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper PT-BR Edresson (Low)

Portuguese BR VITS. 1 speaker. ~64 MB per variant.

Portuguese BR · 1 speaker

Piper VI 25hours (Low)

Vietnamese VITS. 1 speaker. ~64 MB per variant.

Vietnamese · 1 speaker

Piper DE Glados (Medium)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Glados Turret (Medium)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Thorsten (Medium)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper EN-GB Alan (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN-GB Alba (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN-GB Cori (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN-GB Jenny (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Amy (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Bryce (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN HFC Female (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN HFC Male (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Joe (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN John (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Kristin (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Kusal (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Lessac (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN LJSpeech (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Norman (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Reza (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Ryan (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Sam (Medium)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper ES Glados (Medium)

Spanish VITS. 1 speaker. ~64 MB per variant.

Spanish · 1 speaker

Piper ES-MX Claude (High)

Spanish VITS. 1 speaker. ~64 MB per variant.

Spanish · 1 speaker

Piper FA Ganji (Medium)

Persian VITS. 1 speaker. ~64 MB per variant.

Persian · 1 speaker

Piper FA Ganji Adabi (Medium)

Persian VITS. 1 speaker. ~64 MB per variant.

Persian · 1 speaker

Piper FA Gyro (Medium)

Persian VITS. 1 speaker. ~64 MB per variant.

Persian · 1 speaker

Piper FA Reza (Medium)

Persian VITS. 1 speaker. ~64 MB per variant.

Persian · 1 speaker

Piper FR Siwis (Medium)

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper FR Tjiho Model1

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper FR Tjiho Model2

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper FR Tjiho Model3

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper HI Pratham (Medium)

Hindi VITS. 1 speaker. ~64 MB per variant.

Hindi · 1 speaker

Piper HI Priyamvada (Medium)

Hindi VITS. 1 speaker. ~64 MB per variant.

Hindi · 1 speaker

Piper HU Berta (Medium)

Hungarian VITS. 1 speaker. ~64 MB per variant.

Hungarian · 1 speaker

Piper HU Imre (Medium)

Hungarian VITS. 1 speaker. ~64 MB per variant.

Hungarian · 1 speaker

Piper IS Salka (Medium)

Icelandic VITS. 1 speaker. ~64 MB per variant.

Icelandic · 1 speaker

Piper IS Steinn (Medium)

Icelandic VITS. 1 speaker. ~64 MB per variant.

Icelandic · 1 speaker

Piper IS Ugla (Medium)

Icelandic VITS. 1 speaker. ~64 MB per variant.

Icelandic · 1 speaker

Piper ML Arjun (Medium)

Malayalam VITS. 1 speaker. ~64 MB per variant.

Malayalam · 1 speaker

Piper NL Pim (Medium)

Dutch VITS. 1 speaker. ~64 MB per variant.

Dutch · 1 speaker

Piper NL-BE Nathalie (Medium)

Dutch VITS. 1 speaker. ~64 MB per variant.

Dutch · 1 speaker

Piper PL Gosia (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PL Jarvis (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PL Justyna (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PL MC Speech (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PL Meski (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PL Zenski (Medium)

Polish VITS. 1 speaker. ~64 MB per variant.

Polish · 1 speaker

Piper PT-BR Cadu (Medium)

Portuguese BR VITS. 1 speaker. ~64 MB per variant.

Portuguese BR · 1 speaker

Piper PT-BR Jeff (Medium)

Portuguese BR VITS. 1 speaker. ~64 MB per variant.

Portuguese BR · 1 speaker

Piper RU Denis (Medium)

Russian VITS. 1 speaker. ~64 MB per variant.

Russian · 1 speaker

Piper RU Dmitri (Medium)

Russian VITS. 1 speaker. ~64 MB per variant.

Russian · 1 speaker

Piper RU Ruslan (Medium)

Russian VITS. 1 speaker. ~64 MB per variant.

Russian · 1 speaker

Piper SV Lisa (Medium)

Swedish VITS. 1 speaker. ~64 MB per variant.

Swedish · 1 speaker

Piper TR DFKI (Medium)

Turkish VITS. 1 speaker. ~64 MB per variant.

Turkish · 1 speaker

Piper TR Fahrettin (Medium)

Turkish VITS. 1 speaker. ~64 MB per variant.

Turkish · 1 speaker

Piper CY Bu TTS (Medium)

Welsh multi-speaker VITS. ~77 MB per variant.

Welsh · multi-speaker

Piper EN-GB ARU (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN-GB Semaine (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN-GB South Female (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN-GB South Male (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN-GB VCTK (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN Arctic (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper EN L2Arctic (Medium)

English multi-speaker VITS. ~77 MB per variant.

English · multi-speaker

Piper ES Sharvard (Medium)

Spanish multi-speaker VITS. ~77 MB per variant.

Spanish · multi-speaker

Piper FR UPMC (Medium)

French multi-speaker VITS. ~77 MB per variant.

French · multi-speaker

Piper NE Google (Medium)

Nepali multi-speaker VITS. ~77 MB per variant.

Nepali · multi-speaker

Piper AR Dii (High)

Arabic VITS. 1 speaker. ~64 MB per variant.

Arabic · 1 speaker

Piper AR Miro (High)

Arabic VITS. 1 speaker. ~64 MB per variant.

Arabic · 1 speaker

Piper AR Miro V2 (High)

Arabic VITS. 1 speaker. ~64 MB per variant.

Arabic · 1 speaker

Piper DE Dii (High)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper DE Miro (High)

German VITS. 1 speaker. ~64 MB per variant.

German · 1 speaker

Piper EN-GB Dii (High)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN-GB Miro (High)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper EN Miro (High)

English VITS. 1 speaker. ~64 MB per variant.

English · 1 speaker

Piper ES Miro (High)

Spanish VITS. 1 speaker. ~64 MB per variant.

Spanish · 1 speaker

Piper FR Miro (High)

French VITS. 1 speaker. ~64 MB per variant.

French · 1 speaker

Piper IT Dii (High)

Italian VITS. 1 speaker. ~64 MB per variant.

Italian · 1 speaker

Piper IT Miro (High)

Italian VITS. 1 speaker. ~64 MB per variant.

Italian · 1 speaker

Piper NL Dii (High)

Dutch VITS. 1 speaker. ~64 MB per variant.

Dutch · 1 speaker

Piper NL Miro (High)

Dutch VITS. 1 speaker. ~64 MB per variant.

Dutch · 1 speaker

Piper PT-BR Dii (High)

Portuguese BR VITS. 1 speaker. ~64 MB per variant.

Portuguese BR · 1 speaker

Piper PT-BR Miro (High)

Portuguese BR VITS. 1 speaker. ~64 MB per variant.

Portuguese BR · 1 speaker

Piper PT Dii (High)

Portuguese VITS. 1 speaker. ~64 MB per variant.

Portuguese · 1 speaker

Piper PT Miro (High)

Portuguese VITS. 1 speaker. ~64 MB per variant.

Portuguese · 1 speaker

Coqui

Coqui BG (Bulgarian)

Bulgarian VITS. Character-level tokenization. Single speaker.

Bulgarian · 1 speaker

Coqui BN (Bengali)

Bengali VITS. Character-level tokenization. Single speaker.

Bengali · 1 speaker

Coqui CS (Czech)

Czech VITS. Character-level tokenization. Single speaker.

Czech · 1 speaker

Coqui DA (Danish)

Danish VITS. Character-level tokenization. Single speaker.

Danish · 1 speaker

Coqui DE (German)

German VITS. Character-level tokenization. Single speaker.

German · 1 speaker

Coqui EN (LJSpeech)

English VITS (LJSpeech). Character-level tokenization. Single speaker.

English · 1 speaker

Coqui EN (Neon)

English VITS (LJSpeech Neon variant). Character-level tokenization. Single speaker.

English · 1 speaker

Coqui EN (VCTK)

English VITS multi-speaker (VCTK). 109 speakers with espeak-ng.

English · 109 speakers

Coqui ES (Spanish)

Spanish VITS. Character-level tokenization. Single speaker.

Spanish · 1 speaker

Coqui ET (Estonian)

Estonian VITS. Character-level tokenization. Single speaker.

Estonian · 1 speaker

Coqui FI (Finnish)

Finnish VITS. Character-level tokenization. Single speaker.

Finnish · 1 speaker

Coqui FR (French)

French VITS. Character-level tokenization. Single speaker.

French · 1 speaker

Coqui GA (Irish)

Irish VITS. Character-level tokenization. Single speaker.

Irish · 1 speaker

Coqui HR (Croatian)

Croatian VITS. Character-level tokenization. Single speaker.

Croatian · 1 speaker

Coqui LT (Lithuanian)

Lithuanian VITS. Character-level tokenization. Single speaker.

Lithuanian · 1 speaker

Coqui LV (Latvian)

Latvian VITS. Character-level tokenization. Single speaker.

Latvian · 1 speaker

Coqui MT (Maltese)

Maltese VITS. Character-level tokenization. Single speaker.

Maltese · 1 speaker

Coqui NL (Dutch)

Dutch VITS. Character-level tokenization. Single speaker.

Dutch · 1 speaker

Coqui PL (Polish)

Polish VITS. Character-level tokenization. Single speaker.

Polish · 1 speaker

Coqui PT (Portuguese)

Portuguese VITS. Character-level tokenization. Single speaker.

Portuguese · 1 speaker

Coqui RO (Romanian)

Romanian VITS. Character-level tokenization. Single speaker.

Romanian · 1 speaker

Coqui SK (Slovak)

Slovak VITS. Character-level tokenization. Single speaker.

Slovak · 1 speaker

Coqui SL (Slovenian)

Slovenian VITS. Character-level tokenization. Single speaker.

Slovenian · 1 speaker

Coqui SV (Swedish)

Swedish VITS. Character-level tokenization. Single speaker.

Swedish · 1 speaker

Coqui UK (Ukrainian)

Ukrainian VITS. Character-level tokenization. Single speaker.

Ukrainian · 1 speaker

Mimic3

Mimic3 AF (Afrikaans)

Afrikaans VITS with espeak-ng. Single speaker.

Afrikaans · 1 speaker

Mimic3 BN (Bengali)

Bengali VITS with espeak-ng. Single speaker.

Bengali · 1 speaker

Mimic3 EL (Greek)

Greek VITS with espeak-ng. Single speaker.

Greek · 1 speaker

Mimic3 ES (Spanish)

Spanish VITS with espeak-ng. Single speaker.

Spanish · 1 speaker

Mimic3 FA (Persian)

Persian VITS with espeak-ng. Single speaker.

Persian · 1 speaker

Mimic3 FI (Finnish)

Finnish VITS with espeak-ng. Single speaker.

Finnish · 1 speaker

Mimic3 GU (Gujarati)

Gujarati VITS with espeak-ng. Single speaker.

Gujarati · 1 speaker

Mimic3 HU (Hungarian)

Hungarian VITS with espeak-ng. Single speaker.

Hungarian · 1 speaker

Mimic3 KO (Korean)

Korean VITS with espeak-ng. Single speaker.

Korean · 1 speaker

Mimic3 NE (Nepali)

Nepali VITS with espeak-ng. Single speaker.

Nepali · 1 speaker

Mimic3 PL (Polish)

Polish VITS with espeak-ng. Single speaker.

Polish · 1 speaker

Mimic3 TN (Tswana)

Tswana VITS with espeak-ng. Single speaker.

Tswana · 1 speaker

Mimic3 VI (Vietnamese)

Vietnamese VITS with espeak-ng. Single speaker.

Vietnamese · 1 speaker

MMS (Massively Multilingual Speech)

MMS DEU (German)

German MMS VITS. Grapheme-based tokenization. Single speaker.

German · 1 speaker

MMS ENG (English)

English MMS VITS. Grapheme-based tokenization. Single speaker.

English · 1 speaker

MMS FRA (French)

French MMS VITS. Grapheme-based tokenization. Single speaker.

French · 1 speaker

MMS NAN (Min Nan Chinese)

Min Nan Chinese MMS VITS. Grapheme-based tokenization. Single speaker.

Min Nan · 1 speaker

MMS RUS (Russian)

Russian MMS VITS. Grapheme-based tokenization. Single speaker.

Russian · 1 speaker

MMS SPA (Spanish)

Spanish MMS VITS. Grapheme-based tokenization. Single speaker.

Spanish · 1 speaker

MMS THA (Thai)

Thai MMS VITS. Grapheme-based tokenization. Single speaker.

Thai · 1 speaker

MMS UKR (Ukrainian)

Ukrainian MMS VITS. Grapheme-based tokenization. Single speaker.

Ukrainian · 1 speaker

MeloTTS / Cantonese / Icefall / zh-ll (Special VITS)

MeloTTS English

English MeloTTS VITS with lexicon. Single speaker. ~163 MB ONNX.

English · 1 speaker

MeloTTS Chinese-English

Bilingual ZH-EN MeloTTS VITS with lexicon + jieba dict + FSTs. ~163 MB ONNX.

ZH-EN · 1 speaker

Cantonese (xiaomaiiwn)

Cantonese VITS with lexicon + rule FST. ~109 MB ONNX.

Cantonese · 1 speaker

Icefall EN LJSpeech Low

English VITS with espeak-ng. Only ~30 MB ONNX — ultra-fast, smallest English model.

English · 1 speaker

Icefall EN LJSpeech Medium

English VITS with espeak-ng. ~70 MB ONNX — better quality than low variant.

English · 1 speaker

Icefall AISHELL3 (Chinese)

Chinese VITS with lexicon + FSTs. 174 speakers. Only ~29 MB ONNX (fastest Chinese).

Chinese · 174 speakers

Chinese zh-ll

Chinese VITS with lexicon + jieba dict + FSTs. 2 speakers. ~116 MB ONNX.

Chinese · 2 speakers
# Pack models before use:
python3 pack.py kitten                  # Kitten Nano EN v0.2
python3 pack.py kitten-mini             # Kitten Mini EN (8 speakers)
python3 pack.py kokoro                  # Kokoro v1.0 int8
python3 pack.py kokoro-fp32             # Kokoro v1.1 fp32
python3 pack.py matcha-en-ljspeech               # Matcha EN + vocos vocoder
python3 pack.py matcha-zh-baker               # Matcha ZH + vocos vocoder
python3 pack.py zipvoice-distill-int8   # ZipVoice int8 + vocos
python3 pack.py zipvoice-distill-fp32   # ZipVoice fp32 + vocos
python3 pack.py pocket-int8             # Pocket TTS int8
python3 pack.py pocket                  # Pocket TTS fp32
python3 pack.py coqui-en-ljspeech                # Coqui EN (LJSpeech)
python3 pack.py mimic3-ko-kss-low               # Mimic3 Korean
python3 pack.py mms-eng                 # MMS English
python3 pack.py melo-tts-en             # MeloTTS English
python3 pack.py melo-tts-zh-en          # MeloTTS Chinese-English
python3 pack.py cantonese               # Cantonese
python3 pack.py icefall-en-ljspeech-low          # Icefall EN LJSpeech Low (30MB)
python3 pack.py icefall-en-ljspeech-medium       # Icefall EN LJSpeech Medium (70MB)
python3 pack.py icefall-zh-aishell3     # Icefall AISHELL3 (174 speakers)
python3 pack.py zh-ll                   # Chinese zh-ll (2 speakers)
python3 pack.py piper-en-amy-low            # Piper EN Amy Low
python3 pack.py piper-de-thorsten-medium    # Piper DE Thorsten Medium
python3 pack.py piper-en-libritts_r-medium  # Piper EN LibriTTS-R Medium (fp32 + int8)
python3 pack.py all                     # all models, all variants

python3 -m http.server 8080