Metinden Söze
Metini açyk çeşme AI modelleri bilen dogry sesli söze öwür. Mugt ulanmak, hasap gerek däl.
Metini SSML taglarda dolap dogry kontrol üçin:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Göndermek üçin emotion markers ekle (model goldawy üýtgeýär):
Öz sözleriň terjimesini belli et (söz = terjime):
Modeliň Täzelikleri
Qwen3 TTS
Qwen3-TTS is a 1.7 billion parameter text-to-speech model from Alibaba's Qwen team. It supports three modes: preset voices with emotion control (9 speakers), voice cloning from just 3 seconds of audio, and a unique voice design mode where you describe the voice you want in natural language. It covers 10 languages with high expressiveness and natural prosody.
| Programçi: | Alibaba (Qwen) |
| Lisenziýa: | Apache 2.0 |
| Tizlik | Medium |
| _Hili: | |
| dil | 10 dil |
| VRAM | 7GB |
| Ses Klonlama | Saýlawlar |
Iň gowy netijeler üçin maslahatlar
- Takyk aralyklar we intonasiýa üçin dogry aralyklary ullan
- Rakamlary we gysgalamalary has aňsat okamak üçin emle
- Sözler arasyna gysga wagtlyk durmalary döretmek üçin çemçeleri goş
- Uzak dramatik wagtlyk durmalar üçin ellipsis (...) ullan
- Kokoro ýa-da CosyVoice 2'i iň nazik netijeler üçin synla
- Birnäçe sözleýji düwmeleri we podkast mazmuny üçin Dia ullan
Karakter Ullanyşy
| _Göçür | 1K karakter üçin töleg |
|---|---|
| Boş | 0 kredit (süýtgäýmez) |
| _Öň bellenen | 2 kredit / 1K karakter |
| Premium | 4 kredit / 1K karakter |
AI Metinden Söze işleýşi
Üç aňsat ädimde professional-kvalite diller döret. Hiç bir tehnikal bilim gerek däl.
Metini giriz
Söze öwürmek isleýän metiniňizi ýaz, guý, ýa-da ýükläň. Ýatdan çykan ulanyjylar üçin her bir nesil üçin 5000 karaktere çenli goldaýar. Saý metini ullan ýa-da SSML taglary goş, sözlem, wagtlaýyn durma we üns bermek üçin ileri kontrol üçin.
Modeli we Sesni Saýla
Üç derejeden 20+ AI modellerden saýlaň. Siziň mazmunyňyza laýyk gelýän ses saýlaň, maksat dili saýlaň, çalmak tizligini 0.5x-den 2.0x-e düzeltin, we islendik çykdajy formatyny saýlaň (MP3, WAV, OGG, ýa-da FLAC).
Bejer
Bejer düw we seniň audioň sekundlar içinde taýyn bolar. Içe giren çalşyjy bilen öňki görkez, saýlan formatda ýükle, ýa-da paýlaşylan baglanyşy nusgala. API'den batch işlemek we iş akymyna birikmek üçin peýdalan.
Metinden Söze Ullanyş Hadysalary
AI-powered text-to-speech adamlaryň audio mazmun bilen döredijilik, sarp etmek we özara hereket etmek usulyny onlarça pudaklarda üýtgedýär.
Metinden Söze
Her bir AI modeli üçin TTS.aida elýeterli detallar. Siziň proýektiňiz üçin iň gowy modeli tapmak üçin hilini, tizligini, dil goldawyny we aýratynlyklary deňeşdiriň.
Kokoro
Free
Kokoro 82 million parametrli metinden söze modeldir we öz agramy klasyndan has gowy işleýär. Kiçijik ululygyna garamazdan, ol örän nazik we beýanly sözi döredýär. Kokoro birnäçe dilleri goldaýar, şol sanda inglizçe, ýaponça, çinçe we koreýçe, dürli beýanly sesler bilen. Ol gaty tiz işleýär — GPU-da hakykat wagtyndan 100x tiz ses döredýär.
Hexgrad
Apache 2.0
Fast
en, ja, zh, ko, fr, de, it, pt, es, hi, ru
1.5GB
_Ýok
Boş
Piper
Free
Piper VITS we larynx binagärliklerden peýdalanýan Rhasspy tarapyndan işlenip çykarylan ýeňil metinden söze enjamdyr. Ol doly CPU üstünde işleýär, bu ony edge enjamlar, öý awtomatlaşdyrma we off-line TTS talap edýän programler üçin ideal edýär. 30+ dilden 100 ses bilen, Piper Raspberry Pi 4-de hem hakykat wagtyň tizliginde tebigy sesli sözi hödürleýär.
Rhasspy
MIT
Fast
en, de, fr, es, it, pt, nl, pl, ru, zh, ja, ko, ar, cs, da, fi, el, hu, is, ka, kk, ne, no, ro, sk, sr, sv, sw, tr, uk, vi
0 (CPU only)
_Ýok
Boş
VITS
Free
VITS (Variasion Inference with adversarial learning for end-to-end Text-to-Speech) şu wagtky iki tapgyrly modellerden has has natural sounding audio döredýän bir TTS usuly. Ol normalizing flows bilen artdyrylan variasion inference we adversarial training prosesi kabul edip, naturalness-de bir uly gowylaşma gazanýar.
Jaehyeon Kim et al.
MIT
Fast
en, zh, ja, ko
1GB
_Ýok
Boş
MeloTTS
Free
MeloTTS by MyShell.ai bir köp dilli TTS kitaphanasy bolup inglizçe (Amerikan, Britanik, Hindistan, Awstralýa), ispança, fransuzça, çinçe, japonça we koreýçe goldaýar. Ol gaty tiz, metini diňe CPU bilen hakykatdan ykjam tizlikte işleýär. MeloTTS öndürijilik ulanmak üçin işlenip taýýarlanyldy we hem CPU hem GPU inference goldaýar.
MyShell.ai
MIT
Fast
en, es, fr, zh, ja, ko
0.5GB (GPU optional)
_Ýok
Boş
Bark
Standard
Suno tarapyndan Bark bir transformator esaslanan metinden ses modeldir, ol ýokary realistik, köp dilli sözi hemde beýleki sesleri, ýagny müzik, arka fon gürültüsüni we ses effektlerini döredip biler. Ol gülmek, öwmek we gygyrmak ýaly sözsüz aragatnaşygy döredip biler. Bark 100den gowrak sözçi öň bellemeleri we 13den gowrak dili goldaýar.
Suno
MIT
Slow
en, zh, fr, de, hi, it, ja, ko, pl, pt, ru, es, tr
5GB
_Ýok
2x
Bark Small
Standard
Bark Small Bark modeliň bir distilllenmiş wersiýasy bolup, ol belli bir ses hilinden has çalt netije çykarmak tizligi we has az ýat talaplary üçin alyş-çalşyp bilýär. Ol Bark'yň sözi duýgular, gülüşmeler we köp dil bilen döredip bilýänligini saklaýar.
Suno
MIT
Medium
en, zh, fr, de, hi, it, ja, ko, pl, pt, ru, es, tr
2GB
_Ýok
2x
CosyVoice 2
Standard
Alibaba's Tongyi Lab tarapyndan döredilen CosyVoice 2 adama deňeşdirilen söz hilini aňsatlyk bilen gazanýar, ony hakykat wagtynda programmalar üçin ideal edýär. Ol ströming sintez üçin bir sonly skalar kwantlaşdyrma çemeleşmesini ulanýar we zero-shot ses klonlamany, cross-language syntez, we fin-grained emosiýa kontroly goldaýar. Ol subýektiw bahalarda köp tiz TTS sistemalardan üstün çykýar.
Alibaba (Tongyi Lab)
Apache 2.0
Medium
en, zh, ja, ko, fr, de, it, es
4GB
Eý
2x
Dia TTS
Standard
Nari Labs tarapyndan işlenip çykarylan Dia bir 1.6B parametrli metinden söze modeldir, aýratyn bir köp sözleýji dialogy döretmek üçin işlenip çykarylan. Ol iki sözleýji arasyna dogry aýlaw, prozodiya we duýgu beýany bilen dogry sesli söhbetdeşlik döretmäge ukyply. Dia podcast-style mazmun, audiokitap söhbetdeşlik, we interaktiw gürleşýän AI döretmek üçin täsindir.
Nari Labs
Apache 2.0
Medium
en
4GB
_Ýok
2x
Parler TTS
Standard
Parler TTS bir metinden söze modeldir, ol döredilen sözi dolandyrmak üçin tebigy dil ses beýanyny ulanýar. Ön bellenen seslerden saýlamak ýerine, islän sesi beýan ediň (m.a., "a warm female voice with a slight British accent, speaking slowly and clearly") we Parler şol beýana laýyk ses döreder. Bu ony döredijilikli programler üçin özüne mahsus elýeterli edýär.
Hugging Face
Apache 2.0
Medium
en
4GB
_Ýok
2x
GLM-TTS
Standard
GLM-TTS Zhipu AI tarapyndan Llama binagärligi bilen öçüm deňlemeli metin-söz sistemasy. Ol açyk çeşme TTS modelleri arasynda iň az karakter hata derejesine ýetýär, munuň manysy ol iň dogry sözlemleri çykarýar. GLM-TTS 3-10 sekunt ses nusgalaryndan ses klonlama bilen ingliz we çin dilini goldaýar.
Zhipu AI
GLM-4 License
Medium
en, zh
4GB
Eý
2x
IndexTTS-2
Standard
IndexTTS-2 bir ileri metin-söz sistemasy bolup, ol 0-şot ses sintezinde we gowy-gök emosiýa kontrol etmekde üstünlik gazanýar. Ol emosiýa spesifiki tälim bermegiň zerurlygy bolmadyk, şat, gahar, gaharlan ýa-da gorkan ýaly belli emosional tony bilen sözi döredip biler. Model, emotion vektorlary ulanyp, emotion ekspressini dogry kontrol etmek üçin sözi döredýär.
Index Team
Bilibili Model License
Medium
en, zh
4GB
Eý
2x
Spark TTS
Standard
Spark TTS by SparkAudio ses klonlamany kontrol ediji emosiýa we gürleýiş stili bilen birikdirýän bir metinden-söz modeldir. Diňe 5 sekunt referens ses ulanyp, ol ses klonlap soňra dürli emosiýalar, tizlikler we stiller bilen gürleýişi döredip biler. Spark TTS bir soraga esaslanan kontrol sistemany ulanýar.
SparkAudio
CC BY-NC-SA 4.0
Medium
en, zh
4GB
Eý
2x
GPT-SoVITS
Standard
GPT-SoVITS GPT-style dil modellemesini SoVITS (Singing Voice Inference via Translation and Synthesis) bilen birikdirýär güýçli az-şot ses klonlamak üçin. 5 sekuntdan az referenç ses bilen, ol ses klonlap we täze sözi döredip we sözleýjiniň özboluşly häsiýetlerini saklaýar. Ol hem gürlemekde hemem aýdym aýtmakda ses sintez etmekde örän gowy.
RVC-Boss
MIT
Slow
en, zh, ja, ko
6GB
Eý
2x
Orpheus
Standard
Orpheus bir uly ölçegli metinden söze modeldir, ol adam derejesindäki duýgulary beýan eder. 100,000 sagatdan gowrak dürli duýgular maglumaty bilen taýýarlanan, ol dogry duýgular, üns, we duýgu stilleri bilen sözi döredip bilýär. Orpheus adam ýazgylaryndan tapawutly bolmadyk sözi döredip bilýär.
Canopy Labs
Llama 3.2 Community
Medium
en
4GB
_Ýok
2x
Chatterbox
Premium
Resemble AI tarapyndan Chatterbox bir zero-shot ses klonlama modelidir. Ol bir ses nusgasyndan her bir sesi has dogrylyk bilen, diňe timbre däl, ýöne sözleme stili we duýgular bilen hem duýgulary saklaýar. Chatterbox şeýle hem size sesden tapawutly duýgular bilen sözlemeni düzetmäge mümkinçilik berýän duýgular kontroly bilen hem üpjün edýär.
Resemble AI
MIT
Medium
en
4GB
Eý
4x
Tortoise TTS
Premium
Tortoise TTS bir awto-regresiw köp sesli metinden söze sistemasy bolup sesiň hiliniň tizligine öñlük berer. Ol DALL-E-den ilham alan binagärligi gowy prozody we sözleýjiň meňzeşligi bilen ýokary derejede nazik sözi döretmek üçin ulanýar. Birçok alternatiwlerden yavaşrak bolsada, Tortoise azat çeşme ekosistemasynda elýeterli iň hakykatçy sintetik sözi döretýär.
James Betker
Apache 2.0
Slow
en
8GB
Eý
4x
StyleTTS 2
Premium
StyleTTS 2 adam derejeli TTS sintezini, uly söz dil modelleri ulanyp, stil diffýuziya bilen garşydaş tälimni birleşdirmek arkaly gazanýar. Ol bir sözleýji modelleri arasynda iň nazik sesli sözi döredýär, adam ýazgylary bilen bäsleşip. StyleTTS 2 adam sözüni üýtgetmegiň doly spektrini almak üçin diffýuziya esasly stil modellemesini ulanýar.
Columbia University
MIT
Medium
en
4GB
_Ýok
4x
OpenVoice
Premium
MyShell.ai tarapyndan OpenVoice sesiň stili, emosiýa, aksent, ritmi, duralma we intonasiýa üstünden has gowy kontrol bilen sesiň tiz klonlanmagy mümkin edýär. Ol sesiň gysga ses klipinden klonlanmagy we birnäçe dillerde sözi döretmegi mümkin edýär we sözleýjiniň kimligini saklaýar. OpenVoice bir ses konverteri bolup hem işleýär, real wagt ses transformasiýasyny mümkin edýär.
MyShell.ai / MIT
MIT
Medium
en, zh, ja, ko, fr, de, es, it
4GB
Eý
4x
Qwen3 TTS
Standard
Qwen3-TTS Alibaba's Qwen team-den 1.7 milliard parametr metin-dan-söz modeldir. Ol üç mody goldaýar: emotion kontrol bilen öň bellenen sesler (9 sözleýji), ses klonlamak diňe 3 sekunt sesden, we bir özüne mahsus ses dizaýn mody siz isleýän sesi dogry dil bilen beýan edýäňiz. Ol 10 dili ýokary beýan we dogry prozodiya bilen goraýar.
Alibaba (Qwen)
Apache 2.0
Medium
en, zh, ja, ko, de, fr, ru, pt, es, it
7GB
Eý
2x
Sesame CSM
Premium
Sesame CSM (Conversational Speech Model) 1 milliard parametrli modeldir, söhbetçilik üçin aýratyn işlenip düzülipdir. Ol adam söhbetçilikiň tebigy şablonlaryny, şol sanda aýlaw-almak wagtyny, arka kanal jogaplaryny, duýgu reaksiýalaryny we söhbetçilik akymyny modelleşdirýär. CSM sintetik gepçilikden has adam söhbetçilik ýaly ses döredýär.
Sesame
Apache 2.0
Slow
en
8GB
_Ýok
4x
Chatterbox Turbo
Standard
Resemble AI tarapyndan Chatterbox Turbo Chatterbox üçin 350M parametr täzelemedir, 6x real wagt tizligine 200ms-den az wagtlyk gijäni getirýär. Ol [laugh], [cough], we [chuckle] ýaly paralingwistik taglary metinde dogrydan goldaýar. Perth suw markasyny ähli emele gelen seslerde içer.
Resemble AI
MIT
Fast
en
2GB
Eý
2x
Zonos
Standard
Zonos v0.1 Zyphra tarapyndan 1.6B parametrli modeldir, bagt, gahar, keder, gork we şaşka üçin slaiderler bilen gowy-gök emosional kontroly bilen. Ol hem bir Transformator hemem täze SSM (hal-öz-öz-içi model) varianty hödürleýär. 200K+ sagat köp dilli sözleň bilen 10-30 sekuntlyk referans ses klonlama bilen taýýarlandy.
Zyphra
Apache 2.0
Medium
en, ja, zh, fr, de
6GB
Eý
2x
Dia 2
Standard
Dia2 Nari Labs tarapyndan Dia-a 1B we 2B parametrler bilen elýeterli bir stream-first täzelemedir. Ol sesleri ilkinji birnäçe tokenlerden sintez edip başlar, ony hakykat wagty ses agentleri we sözden söze geçirijiler üçin ideal edip goýýar. [S1]/[S2] taglar bilen köp-sözleýji dialogy we (laughs), (coughs) ýaly paralingvistik çyzgylary goldaýar.
Nari Labs
Apache 2.0
Fast
en
4GB
_Ýok
2x
VoxCPM
Standard
VoxCPM 1.5 OpenBMB tarapyndan bir täze tokenizer-free TTS modeldir, ol diskret tokenlerden has dowamly boşlukda işleýär. Ol ýokary-düşünjeli 44.1kHz ses çykarýar, 3-10 sekuntdan zero-shot ses klonlama goldaýar, we paragraflar arasynda birmeňzeşligi saklaýar. Cross-language klonlama size ingliz sesini Çinçe söze ulanmak we tersine.
OpenBMB
Apache 2.0
Fast
en, zh
4GB
Eý
2x
OuteTTS
Free
OuteTTS metinde-söze ukyplary bilen uly dil modellerini giňeldýär we özbaşdak binagärligi saklaýar. Ol llama.cpp (CPU/GPU), Hugging Face Transformers, ExLlamaV2, VLLM, we hatda Transformers.js arkaly brauzer inferences bilen birnäçe arkadaklar goldaýar. JSON ýaly gaýd edilen sözleýji profiller arkaly zero-shot ses klonlama aýratynlyklary bar.
OuteAI
Apache 2.0
Fast
en
2GB
Eý
Boş
TADA
Standard
TADA (Text-Acoustic Dual Alignment) Hume AI tarapyndan Llama 3.2 üstünde bina edilen täze ikinji birikdirme binagärligi arkaly hallusinasiýalary ýok edýän bir täzelikçi TTS modelidir. 1B (Inglizçe) we 3B (çok dil) variantlarda elýeterli, TADA bir RTF 0.09 - 5x çaltlyk bilen deňeşdirilen LLM- esasly TTS modellerinden ýetişdirýär. Ol 700 sekunt audio konteksti goldaýar we standart benchmarkslarda hiç hallusinasiýalar bilen duýguly beýan eden sözi öndürýär.
Hume AI
MIT
Fast
en
5GB
_Ýok
2x
VibeVoice
Standard
Microsoft VibeVoice iki görnüşinde gelýär: 1.5B model uzak mazmun üçin (90 minuta çenli, 4 sözleýji) we REALTIME 0.5B model ~200ms ilkinji ses geçilişi bilen ströming üçin. 1.5B variant podkastlar we audiokitaplarda uzun sözleýji geçişleri bilen üstünlik gazanýar. Ýörelge: Microsoft TTS koduny repozytoriýadan aýyrdy we döreden ses eşidilýän AI jogapkärçilikden gaça durmagy içerýär.
Microsoft
MIT
Fast
en, zh
4GB
_Ýok
2x
Pocket TTS
Free
Kyutai (Moshi-niň döredijileri) tarapyndan Pocket TTS bir 100M parametrli metinde-söze modeldir. Ol CPU-da çalt işleýär, bir ses nusgasyndan zero-shot ses klonlamany goldaýar, we dogry sesli sözi döredýär. Kiçi model ululygy ony edge depleniş we az-resursly töwerekler üçin ideal edýär.
Kyutai
MIT
Fast
en, fr
1GB
Eý
Boş
Kitten TTS
Free
Kitten TTS by KittenML is an ultra-lightweight text-to-speech model built on ONNX. With variants from 15M to 80M parameters (25-80 MB on disk), it delivers high-quality voice synthesis on CPU without requiring a GPU. Features 8 built-in voices, adjustable speech speed, and built-in text preprocessing for numbers, currencies, and units. Ideal for edge deployment and low-latency applications.
KittenML
Apache 2.0
Fast
en
0GB
_Ýok
Boş
CosyVoice3
Standard
CosyVoice3 is the latest evolution from Alibaba's FunAudioLLM team. It features bi-streaming inference with ~150ms latency, instruction-based control for emotion/speed/volume, and improved speaker similarity for zero-shot cloning. Supports 9 languages plus 18 Chinese dialects. RL-tuned variant delivers state-of-the-art prosody.
Alibaba (FunAudioLLM)
Apache 2.0
Fast
en, zh, ja, ko, de, es, fr, it, ru
4GB
Eý
2x
MOSS-TTS
Premium
MOSS-TTS from OpenMOSS supports generation of up to 1 hour of continuous speech across 20 languages. Features token-level duration control, phoneme-level pronunciation control via IPA/Pinyin, and code-switching between languages. The 8B production model delivers state-of-the-art quality with zero-shot voice cloning from reference audio.
OpenMOSS
Apache 2.0
Medium
en, zh, de, es, fr, ja, it, hu, ko, ru, fa, ar, pl, pt, cs, da, sv, el, tr
16GB
Eý
4x
MegaTTS3
Premium
MegaTTS3 from ByteDance uses a novel sparse alignment mechanism combined with a latent diffusion transformer. Features adjustable trade-off between speech intelligibility and speaker similarity for zero-shot voice cloning.
ByteDance
Apache 2.0
Slow
en, zh
8GB
Eý
4x
Kokoro
Boş
Kokoro is an 82 million parameter text-to-speech model that punches well above its weight class. Despite its tiny size, it produces remarkably natural and expressive speech. Kokoro supports multiple languages including English, Japanese, Chinese, and Korean with a variety of expressive voices. It runs incredibly fast — generating audio nearly 100x faster than real-time on a GPU.
Hexgrad
Apache 2.0
Fast
Piper
Boş
Piper is a lightweight text-to-speech engine developed by Rhasspy that uses VITS and larynx architectures. It runs entirely on CPU, making it ideal for edge devices, home automation, and applications requiring offline TTS. With over 100 voices across 30+ languages, Piper delivers natural-sounding speech at real-time speeds even on a Raspberry Pi 4.
Rhasspy
MIT
Fast
VITS
Boş
VITS (Variational Inference with adversarial learning for end-to-end Text-to-Speech) is a parallel end-to-end TTS method that generates more natural sounding audio than current two-stage models. It adopts variational inference augmented with normalizing flows and an adversarial training process, achieving a significant improvement in naturalness.
Jaehyeon Kim et al.
MIT
Fast
MeloTTS
Boş
MeloTTS by MyShell.ai is a multilingual TTS library supporting English (American, British, Indian, Australian), Spanish, French, Chinese, Japanese, and Korean. It is extremely fast, processing text at near real-time speed on CPU alone. MeloTTS is designed for production use and supports both CPU and GPU inference.
MyShell.ai
MIT
Fast
OuteTTS
Boş
OuteTTS extends large language models with text-to-speech capabilities while preserving the original architecture. It supports multiple backends including llama.cpp (CPU/GPU), Hugging Face Transformers, ExLlamaV2, VLLM, and even browser inference via Transformers.js. Features zero-shot voice cloning through speaker profiles saved as JSON.
OuteAI
Apache 2.0
Fast
Pocket TTS
Boş
Pocket TTS by Kyutai (creators of Moshi) is a compact 100M parameter text-to-speech model that punches well above its weight. It runs efficiently on CPU, supports zero-shot voice cloning from a single audio sample, and produces natural-sounding speech. The small model size makes it ideal for edge deployment and low-resource environments.
Kyutai
MIT
Fast
Kitten TTS
Boş
Kitten TTS by KittenML is an ultra-lightweight text-to-speech model built on ONNX. With variants from 15M to 80M parameters (25-80 MB on disk), it delivers high-quality voice synthesis on CPU without requiring a GPU. Features 8 built-in voices, adjustable speech speed, and built-in text preprocessing for numbers, currencies, and units. Ideal for edge deployment and low-latency applications.
KittenML
Apache 2.0
Fast
Bark
_Öň bellenen
Bark by Suno is a transformer-based text-to-audio model that can generate highly realistic, multilingual speech as well as other audio like music, background noise, and sound effects. It can produce nonverbal communications like laughing, sighing, and crying. Bark supports over 100 speaker presets and 13+ languages.
Suno
MIT
Slow
en, zh, fr, de, hi, it, ja, ko, pl, pt, ru, es, tr
_Ýok
Bark Small
_Öň bellenen
Bark Small is a distilled version of the Bark model that trades some audio quality for significantly faster inference speeds and lower memory requirements. It retains Bark's ability to generate speech with emotions, laughter, and multiple languages.
Suno
MIT
Medium
en, zh, fr, de, hi, it, ja, ko, pl, pt, ru, es, tr
_Ýok
CosyVoice 2
_Öň bellenen
CosyVoice 2 by Alibaba's Tongyi Lab achieves human-comparable speech quality with extremely low latency, making it ideal for real-time applications. It uses a finite scalar quantization approach for streaming synthesis and supports zero-shot voice cloning, cross-lingual synthesis, and fine-grained emotion control. It outperforms many commercial TTS systems in subjective evaluations.
Alibaba (Tongyi Lab)
Apache 2.0
Medium
en, zh, ja, ko, fr, de, it, es
Eý
Dia TTS
_Öň bellenen
Dia by Nari Labs is a 1.6B parameter text-to-speech model designed specifically for generating multi-speaker dialogue. It can produce natural-sounding conversations between two speakers with appropriate turn-taking, prosody, and emotional expression. Dia is perfect for creating podcast-style content, audiobook dialogues, and interactive conversational AI.
Nari Labs
Apache 2.0
Medium
en
_Ýok
Parler TTS
_Öň bellenen
Parler TTS is a text-to-speech model that uses natural language voice descriptions to control the generated speech. Instead of selecting from preset voices, you describe the voice you want (e.g., "a warm female voice with a slight British accent, speaking slowly and clearly") and Parler generates speech matching that description. This makes it uniquely flexible for creative applications.
Hugging Face
Apache 2.0
Medium
en
_Ýok
GLM-TTS
_Öň bellenen
GLM-TTS by Zhipu AI is a text-to-speech system built on the Llama architecture with flow matching. It achieves the lowest character error rate among open-source TTS models, meaning it produces the most accurate pronunciation. GLM-TTS supports English and Chinese with voice cloning from 3-10 second audio samples.
Zhipu AI
GLM-4 License
Medium
en, zh
Eý
IndexTTS-2
_Öň bellenen
IndexTTS-2 is an advanced text-to-speech system that excels at zero-shot voice synthesis with fine-grained emotion control. It can generate speech with specific emotional tones like happy, sad, angry, or fearful without requiring emotion-specific training data. The model uses emotion vectors to precisely control the emotional expression of generated speech.
Index Team
Bilibili Model License
Medium
en, zh
Eý
Spark TTS
_Öň bellenen
Spark TTS by SparkAudio is a text-to-speech model that combines voice cloning with controllable emotion and speaking style. Using just 5 seconds of reference audio, it can clone a voice and then generate speech with different emotions, speeds, and styles while maintaining the cloned voice identity. Spark TTS uses a prompt-based control system.
SparkAudio
CC BY-NC-SA 4.0
Medium
en, zh
Eý
GPT-SoVITS
_Öň bellenen
GPT-SoVITS combines GPT-style language modeling with SoVITS (Singing Voice Inference via Translation and Synthesis) for powerful few-shot voice cloning. With as little as 5 seconds of reference audio, it can accurately clone a voice and generate new speech while preserving the speaker's unique characteristics. It excels at both speaking and singing voice synthesis.
RVC-Boss
MIT
Slow
en, zh, ja, ko
Eý
Orpheus
_Öň bellenen
Orpheus is a large-scale text-to-speech model that achieves human-level emotional expression. Trained on over 100,000 hours of diverse speech data, it excels at generating speech with natural emotions, emphasis, and speaking styles. Orpheus can produce speech that is virtually indistinguishable from human recordings.
Canopy Labs
Llama 3.2 Community
Medium
en
_Ýok
Qwen3 TTS
_Öň bellenen
Qwen3-TTS is a 1.7 billion parameter text-to-speech model from Alibaba's Qwen team. It supports three modes: preset voices with emotion control (9 speakers), voice cloning from just 3 seconds of audio, and a unique voice design mode where you describe the voice you want in natural language. It covers 10 languages with high expressiveness and natural prosody.
Alibaba (Qwen)
Apache 2.0
Medium
en, zh, ja, ko, de, fr, ru, pt, es, it
Eý
Chatterbox Turbo
_Öň bellenen
Chatterbox Turbo by Resemble AI is a 350M parameter upgrade to Chatterbox, delivering up to 6x real-time speed with sub-200ms latency. It supports paralinguistic tags like [laugh], [cough], and [chuckle] directly in text. Includes Perth watermarking on all generated audio for provenance tracking.
Resemble AI
MIT
Fast
en
Eý
Zonos
_Öň bellenen
Zonos v0.1 by Zyphra is a 1.6B parameter model featuring fine-grained emotion control with sliders for happiness, anger, sadness, fear, and surprise. It offers both a Transformer and a novel SSM (state-space model) variant. Trained on 200K+ hours of multilingual speech with zero-shot voice cloning from 10-30 seconds of reference audio.
Zyphra
Apache 2.0
Medium
en, ja, zh, fr, de
Eý
Dia 2
_Öň bellenen
Dia2 by Nari Labs is a streaming-first upgrade to Dia, available in 1B and 2B parameter variants. It begins synthesizing audio from the first few tokens, making it ideal for real-time voice agents and speech-to-speech pipelines. Supports multi-speaker dialogue with [S1]/[S2] tags and paralinguistic cues like (laughs), (coughs).
Nari Labs
Apache 2.0
Fast
en
_Ýok
VoxCPM
_Öň bellenen
VoxCPM 1.5 by OpenBMB is a novel tokenizer-free TTS model that operates in continuous space rather than discrete tokens. It produces high-fidelity 44.1kHz audio, supports zero-shot voice cloning from 3-10 seconds, and maintains consistency across paragraphs. Cross-language cloning lets you apply an English voice to Chinese speech and vice versa.
OpenBMB
Apache 2.0
Fast
en, zh
Eý
TADA
_Öň bellenen
TADA (Text-Acoustic Dual Alignment) by Hume AI is a groundbreaking TTS model that eliminates hallucinations through a novel dual alignment architecture built on Llama 3.2. Available in 1B (English) and 3B (multilingual) variants, TADA achieves an RTF of 0.09 — 5x faster than comparable LLM-based TTS models. It supports up to 700 seconds of audio context and produces emotionally expressive speech with zero hallucinations on standard benchmarks.
Hume AI
MIT
Fast
en
_Ýok
VibeVoice
_Öň bellenen
VibeVoice from Microsoft generates long-form speech up to 90 minutes with support for 4 simultaneous speakers, making it ideal for podcasts and dialogues. The Realtime 0.5B variant achieves ~300ms latency for interactive use. Supports speaker tags for multi-turn dialogue generation.
Microsoft
MIT
Fast
en, zh
_Ýok
CosyVoice3
_Öň bellenen
CosyVoice3 is the latest evolution from Alibaba's FunAudioLLM team. It features bi-streaming inference with ~150ms latency, instruction-based control for emotion/speed/volume, and improved speaker similarity for zero-shot cloning. Supports 9 languages plus 18 Chinese dialects. RL-tuned variant delivers state-of-the-art prosody.
Alibaba (FunAudioLLM)
Apache 2.0
Fast
en, zh, ja, ko, de, es, fr, it, ru
Eý
Model deňleşik täblisi
| Mody | Programçi: | _Göçür | _Hili: | Tizlik | dil | Ses Klonlama | VRAM | Lisenziýa: | kredit | |
|---|---|---|---|---|---|---|---|---|---|---|
| Kokoro | Hexgrad | Free | Fast | 11 | 1.5GB | Apache 2.0 | Boş | Ullan | ||
| Piper | Rhasspy | Free | Fast | 31 | 0 (CPU only) | MIT | Boş | Ullan | ||
| VITS | Jaehyeon Kim et al. | Free | Fast | 4 | 1GB | MIT | Boş | Ullan | ||
| MeloTTS | MyShell.ai | Free | Fast | 6 | 0.5GB (GPU optional) | MIT | Boş | Ullan | ||
| Bark | Suno | Standard | Slow | 13 | 5GB | MIT | 2 | Ullan | ||
| Bark Small | Suno | Standard | Medium | 13 | 2GB | MIT | 2 | Ullan | ||
| CosyVoice 2 | Alibaba (Tongyi Lab) | Standard | Medium | 8 | 4GB | Apache 2.0 | 2 | Ullan | ||
| Dia TTS | Nari Labs | Standard | Medium | 1 | 4GB | Apache 2.0 | 2 | Ullan | ||
| Parler TTS | Hugging Face | Standard | Medium | 1 | 4GB | Apache 2.0 | 2 | Ullan | ||
| GLM-TTS | Zhipu AI | Standard | Medium | 2 | 4GB | GLM-4 License | 2 | Ullan | ||
| IndexTTS-2 | Index Team | Standard | Medium | 2 | 4GB | Bilibili Model License | 2 | Ullan | ||
| Spark TTS | SparkAudio | Standard | Medium | 2 | 4GB | CC BY-NC-SA 4.0 | 2 | Ullan | ||
| GPT-SoVITS | RVC-Boss | Standard | Slow | 4 | 6GB | MIT | 2 | Ullan | ||
| Orpheus | Canopy Labs | Standard | Medium | 1 | 4GB | Llama 3.2 Community | 2 | Ullan | ||
| Chatterbox | Resemble AI | Premium | Medium | 1 | 4GB | MIT | 4 | Ullan | ||
| Tortoise TTS | James Betker | Premium | Slow | 1 | 8GB | Apache 2.0 | 4 | Ullan | ||
| StyleTTS 2 | Columbia University | Premium | Medium | 1 | 4GB | MIT | 4 | Ullan | ||
| OpenVoice | MyShell.ai / MIT | Premium | Medium | 8 | 4GB | MIT | 4 | Ullan | ||
| Qwen3 TTS | Alibaba (Qwen) | Standard | Medium | 10 | 7GB | Apache 2.0 | 2 | Ullan | ||
| Sesame CSM | Sesame | Premium | Slow | 1 | 8GB | Apache 2.0 | 4 | Ullan | ||
| Chatterbox Turbo | Resemble AI | Standard | Fast | 1 | 2GB | MIT | 2 | Ullan | ||
| Zonos | Zyphra | Standard | Medium | 5 | 6GB | Apache 2.0 | 2 | Ullan | ||
| Dia 2 | Nari Labs | Standard | Fast | 1 | 4GB | Apache 2.0 | 2 | Ullan | ||
| VoxCPM | OpenBMB | Standard | Fast | 2 | 4GB | Apache 2.0 | 2 | Ullan | ||
| OuteTTS | OuteAI | Free | Fast | 1 | 2GB | Apache 2.0 | Boş | Ullan | ||
| TADA | Hume AI | Standard | Fast | 1 | 5GB | MIT | 2 | Ullan | ||
| VibeVoice | Microsoft | Standard | Fast | 2 | 4GB | MIT | 2 | Ullan | ||
| Pocket TTS | Kyutai | Free | Fast | 2 | 1GB | MIT | Boş | Ullan | ||
| Kitten TTS | KittenML | Free | Fast | 1 | 0GB | Apache 2.0 | Boş | Ullan | ||
| CosyVoice3 | Alibaba (FunAudioLLM) | Standard | Fast | 9 | 4GB | Apache 2.0 | 2 | Ullan | ||
| MOSS-TTS | OpenMOSS | Premium | Medium | 19 | 16GB | Apache 2.0 | 4 | Ullan | ||
| MegaTTS3 | ByteDance | Premium | Slow | 2 | 8GB | Apache 2.0 | 4 | Ullan |
Ençeme çeper metinde ses platformy
Metinden Söze Näme Uçin TTS.ai Saýlaň?
TTS.ai dünýäniň iň gowy açyk çeşme metinden söze modellerini bir, ulanmak aňsat platformada birikdirýär. Siziň bir ses enjininde gaplanýan hususy hyzmatlardan tapawutly TTS.ai size Coqui, MyShell, Amphion, NVIDIA, Suno, HuggingFace, Tsinghua University we başgalar ýaly öňdebaryjy gözleg laboratoriýalardan 20+ modellere elýeterliligi berýär.
Her bir model MIT, Apache 2.0 ýa-da şonuň ýaly ygtybarly lisensiýalar astynda açyk çeşmedir, siziň proýektiňizde öndürilen sesleri ulanmak üçin doly söwda hukugyňyzy kepillendirýär. Siz tiz, ýeňil sintez üçin hakykatdanam wagtyň programmalary ýa-da premium studiýa-hilli çykdajy üçin audiokitaplar we podkastlar isleseňiz, TTS.ai her bir ulanmak ýagdaýy üçin dogry modeldir.
Beýiklik
Üç azat TTS mody bilen derhal başlaň: Piper (ultra-gyzykly, ýeňil), VITS (ýokary hilli neural sintez), we MeloTTS (çok dil goldawy). Başga hiç zat gerek däl, kredit kart gerek däl, we nesillerde çäklendirme ýok. Azat mody inglizçe we başga birçok dilleri goldaýar, aňsat sesli çykdajy bilen, köp programmalar üçin laýyk.
GPU-Tizlediş
Tüm TTS modelleri tiz, dogry emele gelş wagty üçin NVIDIA GPUs üstünde işleýär. Beýik modeller 2 sekuntdan az wagtyň içinde ses emele getirýär. Kokoro, CosyVoice 2, we Bark ýaly standart modeller ortaça 3-5 sekuntda. Tortoise we Chatterbox ýaly iň ýokary hilli premium modeller metin uzynlygyna bagly 5-15 sekuntda işlemeýär.
30+ dil goldanylýar
30 dilden gowrak dilde ses döret, şol sanda inglizçe, ispança, fransuzça, almança, italiýança, portugalça, çinçe, japonça, koreýçe, arapça, hindiçe, rusça we başgalar. Birnäçe modeller dilleriň arasynda sintezleri goldaýar, bu siziň dilleriň arasynda ses döretmäge mümkinçilik berýär. CosyVoice 2 we GPT-SoVITS dilleriň arasynda ses klonlamada üstünlik gazanýar.
Öňki
TTS.ai-i öz programmaňyza biziň OpenAI-a laýyk REST API bilen birleşdiriň. 20+ model üçin bir ahtar. Python, JavaScript, cURL, we Go SDKs. Hakyky wagt programmalar üçin stream goldawy. uly ölçegli mazmun döretmek üçin batch işlemek. async habarlaşmalar üçin Webhooks. Pro we Enterprise planlarda elýeterli.
Gynançly Soraglar
Biz nämeni gowy edip bileris? Siziň pikiriňiz bize kynçylyklary düzetmäge kömek eder.
Metinden Söze Döwme
TTS.ai ulanyp müňlerçe döredijilere goşulyşyň. Täze hasabyňyz bilen 15,000 beýik karakterleri alyň. Beýik modeller ýazgysyz elýeterli.