StyleTTS 2
Free.ai (self-hosted)
·
tts
·
~500 tokens per clip
·
4.6
ka 5 user-ka-qaabka-kan
StyleTTS 2 waa a Codka qoraalka-u-hadalka dhisay by Yinghao Aaron Li. ugu xoog badan ee Voice cloning without training data; expressive multilingual speech.. Is-hoosaysiinta Free.ai GPUs - bilaash ayaa ka socda ishaada maalinlaha ah (500 tokens mid kasta). Soo saaray hoos MIT - isticmaalka ganacsi ee oggolaan Free.ai.
Models la mid ah
Su'aalaha badanaa la isweydiiyo
StyleTTS 2 taageeraa noocyo kala duwan oo afaf ah. Liiska saxda ah ku xiran yahay mashiinka; foomka on page this aqbalaa qoraal kasta oo mashiinka ku soo bandhigi doonaa luqadaha ay taageerayaan. eeg / codka / si ay u buuxaan multi-engine dooran haddii aad u baahan tahay af gaar ah.
Inta badan mashiinada soo bandhigaan neutral-American Ingiriisi default iyo gobol-ku habboon ee afka luqadaha aan Ingiriisi ahayn. Premium mashiinada waxay u muuqdaan in ay kala duwanaansho afka - paste tusaale in la barbar dhigo.
SSML taageerada kala duwan yahay by engine. Pause, prosody, iyo tags xusuusin waxaa lagu maamuusay on engines premium badan iyo qaar ka mid ah qaar ka mid ah kuwa self-hosted. qoraalka caadiga ah waa in si joogto ah u shaqeeyaan - markup la'aan loo baahan yahay.
Streaming TTS waa la heli karaa on premium engines via the /v1/tts/ API endpoint with stream = true. Web UI on this page returns the full clip once rendering finishes.
StyleTTS 2 ku socda GPUs our gaar ah. Generation ka soo qaadaa aad maalin kasta free pool hore. Marka la soo afjaro, lacagta lagu bixiyo tokens bilaabaan $ 5 → 200,000 tokens. Dhab ahaan ~ 5 tokens character kasta, ugu yaraan 100 clip kasta.
Inta u dhaxaysa 5,000 xaraf oo codsi kasta oo ku saabsan web UI. For qaybo dheer (audiobooks, qaybaha buuxa), isticmaal / codka / audiobook / oo chunks iyo jilicsan oo otomaatig ah, ama u yeedha API in loop.
Haa - POST liiska strings in /v1/tts/batch /, ama isticmaalka shaqada UI at /workspace / in ay TTS chain in biyo xireenka dheer (tusaale ahaan, beddel → hadal → xiir).
Haa - POST qoraalka in /v1/tts / oo leh model="StyleTTS 2" (ama slug on page this). Waa in ay ku soo laabtaan WAV ama MP3. Eeg /api/ si aad u eegto buuxa + SDK snippets.
Boggaan waa qoraal-u-hadalka, ma aha isku-dhafka codka - codka waa nidaamka ugu horreeya ee mashiinka. Si aad u abuurto codka (ugu soo dejisan audio soo jeedinta), eeg /voice/clone/, taas oo aad u baahan tahay inaad ama leedahay xuquuqda codka ama aad leedahay oggolaansho qoraal ah oo cad.
Injiillada Self-hosted socdaan Free.ai-oo leh GPUs; waxba ma ka tagaan servers our. Premium engines text u gudbiyaan upstream model bixiyaha hoos DPA our. Waxaanan tababarka ku siin on your inputs oo aan la iibsan xogta.
Haa - Free.ai siin adeegsiga ganacsi ee audio soo saaro. License engine ee hoos (Apache 2.0, MIT, ama shuruudaha iibiyaha) waxaa lagu muujiyey kor iyo on page tusaale ah; in dhaqanka waxaa loola jeedaa voiceovers, ads, podcasts, iyo barnaamijyada oo dhan waa in-qaab.
Haa - shaqo ku fashilantay auto-soo celinta in ay asalka (dhaqameed maalin kasta ama tokens la bixiyo). Haddii la soo celiyo ma muujiyaan maalintii isku mid ah, email contact@free.ai.