Ulwazi-mbhalo lwe-API

Yenza i-Speech to Speech AI ibe yingxenye yezinhlelo zakho zokusebenza nge-REST API elula.

Qala ngokushesha

Qala nge-Speech to Speech AI API ngaphansi kwemizuzu emi-5.

Isisekelo URL
https://api.speechtospeechai.com/v1
1. Start a job

Yonke into i POST ngesinqumo sakho se-API. Isinqumo siyi-id yomsebenzi, hhayi umsindo.

cURL
curl -X POST https://api.speechtospeechai.com/v1/tts/ \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "text=Hello! Welcome to Speech to Speech AI." \
  -F "model=kokoro" \
  -F "voice_id=af_heart" \
  -F "format=mp3"

{"uuid": "3f2c...", "job_id": "...", "status": "queued"}
2. Ukhetho lwemiphumela

Buza umsebenzi ngamasekondi ambalwa kuze kube yilapho isimo siqediwe, bese ulanda i-result_url.

cURL
curl "https://api.speechtospeechai.com/v1/speech/results/?uuid=3f2c..." \
  -H "Authorization: Bearer YOUR_API_KEY"

{"status": "completed", "result_url": "https://.../output.mp3"}

Ukuqinisekisa

Zonke izicelo ze-API zidinga ukuqinisekiswa nge-API key ku- Authorization okubhalwe emantla ephepha.

Okubhalwe emantla ephepha
Authorization: Bearer YOUR_API_KEY
Thola isithonjana sakho se-API kusuka ku- i-akhawunti ye-dashboard.

Ukuklonywa kwezwi

POST /v1/voice-clone/

Ukhuluma umbhalo ngesibizo se-reference recording encane.

Isiqu sombuzo (isiqephu esiningi/ifomu-data)
Amapharamitha Uhlobo Kudingeka Ulwaziso
reference_audio file Yebo Ukurekhoda umsindo ofuna ukuklonywa (10-30 imizuzwana evunyelwe, kuze kube yi-60)
text string Yebo Umbhalo okhuluma ngalo msindo
model string Akukho ibhokisi lokuxoxa, openvoice, tortoise noma gpt-sovits
language string Akukho Ikhodi lesilimi (isiphutha: en)
Uphendulelo

I-id yomsebenzi (uuid). Ucwaningo GET /v1/speech/results/?uuid= kuze kube isimo siqediwe bese ulanda i-result_url.

Umbhalo kuya kumazwi

POST /v1/tts/

Guqula umbhalo ube ingoma yokukhuluma.

Isiqu sombuzo (isiqephu esiningi/ifomu-data)
Amapharamitha Uhlobo Kudingeka Ulwaziso
text string Yebo Umbhalo oguqulwe (max 5000 characters)
model string Akukho kokoro (iphutha), piper, vits noma bark
voice_id string Akukho I-id yomsindo kusuka ku-GET /v1/speech/voices/?model=
language string Akukho Ikhodi lesilimi (isiphutha: en)
speed float Akukho Isiphinda-phindo sejubane (0.5-2.0, okuzenzakalelayo 1.0)
format string Akukho wav (iphutha), mp3, ogg noma flac. Amathegi we-SSML asuswa ngaphambi kokuhlanganisa.
Uphendulelo

I-id yomsebenzi (uuid). Ucwaningo GET /v1/speech/results/?uuid= kuze kube isimo siqediwe bese ulanda i-result_url.

Ukukhuluma ku-Text

POST /v1/transcribe/

Bhala umsindo ube umbhalo osebenzisa ama-timestamps.

Isiqu sombuzo (isiqephu esiningi/ifomu-data)
Amapharamitha Uhlobo Kudingeka Ulwaziso
file file Yebo Ihele lomsindo/levidiyo elizoguqulwa
model string Akukho mkhulu-v3 (iphutha) noma turbo
language string Akukho Ikhodi lesilimi, noma ukuhlola ngokuzenzakalela (isimiso)
enable_word_timestamps bool Akukho Engeza isikhathi sokuqala nokuphela kwegama ngalinye
callback_url string Akukho I-URL ethola imiphumela nge-POST uma umsebenzi uphele
Uphendulelo

I-id yomsebenzi (uuid). Umbuzo GET /v1/transcribe/results/?uuid= kuze kube yilapho isimo siqediwe; impendulo ifaka umbhalo, ulwimi kanye nezingxenye.

Amamodeli atholakalayo

Amamodeli ombhalo-kuya-kwezwi
Imodeli ID Ulwaziso Okungcono kakhulu
kokoro I-TTS ehamba phambili eminingi ulwimi Ukusetshenziswa okujwayelekile, ukuphuma okuphansi
piper TTS elula eminingi ulwimi (amazwi avunyelwe ngokuhweba kuphela) I-Fast renders nge-11 languages
vits I-LJSpeech kanye nabangani base-English be-VCTK abangu-108 IsiNgisi
bark TTS ekhiqizayo, izilungiselelo ezingaphambili ezingu-130 ngezilimi ezingu-13 Ukulungiswa, ukuthunyelwa okuphawulekayo (okusheshayo)
openvoice Ukuklonya umsindo kanye nokushintsha umsindo okukhethwa kukho Isixazululo sebhokisi lezingxoxo
chatterbox Ukulungiswa kwezwi okunge-zero-shot kulezi zinhlobo zezilimi eziyi-23; imodeli yokushintshana kwezwi ejwayelekile Ukuklonya kusuka kusampula encane
Ukukhuluma-ngokubhaliwe
Imodeli ID Ulwaziso Okungcono kakhulu
whisper-large-v3 OpenAI Whisper Large v3 Ukunemba okuphezulu, ulwimi oluningi
whisper-turbo Isikhalazo esithuthukisiwe sokuhamba Ukudluliswa okukhawulelwe

Amakhowudi wephutha

Ikhodi Isimo Ulwaziso
invalid_api_key 401 Isithonjana se-API esingasebenzi noma esilahlekile
insufficient_credits 402 Akukho lutho olunenzuzo olunembile kulolu hlelo
rate_limit_exceeded 429 Izicelo eziningi kakhulu
invalid_request 400 Isifiso esifile
model_not_found 404 Imodeli ecelwe ayikho
voice_not_found 404 Umsindo ocelwe awukho
file_too_large 413 Ihele elilayishiwe lidlula umkhawulo wobukhulu
server_error 500 Iphutha leseva