Uxwebhu lwe API
I-Speech to Speech AI idibanisa kwinkqubo yakho nge-REST API elula.
Isiqalo esikhawulezayo
Qala nge Speech to Speech AI API ngaphantsi kwemizuzu emi-5.
Isiseko se URL
https://api.speechtospeechai.com/v1
1. Qala umsebenzi
Isixhobo ngasinye si-POST ngesitshixo sakho se-API. Impendulo yi-id yomsebenzi, hayi isandi.
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 lweziphumo
Buza umsebenzi ngemizuzwana emininzi de kube iimeko zigqityiwe, emva koko zulazisa i-result_url.
curl "https://api.speechtospeechai.com/v1/speech/results/?uuid=3f2c..." \
-H "Authorization: Bearer YOUR_API_KEY"
{"status": "completed", "result_url": "https://.../output.mp3"}
Uqinisekiso
Zonke izicelo ze-API zifuna uqinisekiso ngesixhobo se-API Authorization isihloko.
Authorization: Bearer YOUR_API_KEY
Upelo lwe Voice
/v1/voice-clone/
Ukuthetha umbhalo ngelizwi le-recording efutshane yokubonisa.
Isiqu sesicelo (inxalenye eninzi/ifomu-data)
| Iparamitha | Udidi | Ifunekayo | Iinketho ze projekti |
|---|---|---|---|
reference_audio |
file | Ewe | Ushicilelo lwesandi sokwenza uclone (10-30 imizuzwana ecetyiswayo, ukuya kuthi ga kwi-60) |
text |
string | Ewe | Umbhalo oza kuthetha ngalo msindo |
model |
string | Akukho nanye | ibhokisi yencoko, openvoice, tortoise okanye gpt-sovits |
language |
string | Akukho nanye | Ikhowudi yelwimi (emiselweyo: en) |
I-response
I-id yomsebenzi (uuid). I-Poll GET /v1/speech/results/?uuid= de i-status igqitywe kwaye ulayishe i-result_url.
Umbhalo ukuya kuSpeech
/v1/tts/
Convert text to speech audio.
Isiqu sesicelo (inxalenye eninzi/ifomu-data)
| Iparamitha | Udidi | Ifunekayo | Iinketho ze projekti |
|---|---|---|---|
text |
string | Ewe | Umbhalo oza kuguqulwa (max 5000 characters) |
model |
string | Akukho nanye | kokoro (okungagqibekanga), piper, vits or bark |
voice_id |
string | Akukho nanye | I-ID yelizwi ukusuka kwi-GET /v1/speech/voices/?model= |
language |
string | Akukho nanye | Ikhowudi yelwimi (emiselweyo: en) |
speed |
float | Akukho nanye | Isixa-mali sesantya (0.5-2.0, okwendalo 1.0) |
format |
string | Akukho nanye | wav (okwendalo), mp3, ogg okanye flac. Ii-tags ze-SSML zisuswa phambi kokwenziwa. |
I-response
I-id yomsebenzi (uuid). I-Poll GET /v1/speech/results/?uuid= de i-status igqitywe kwaye ulayishe i-result_url.
Ukuthetha- ukuthethana
/v1/transcribe/
Uguqulelo lwesandi kumbhalo oneziqinisekiso zexesha.
Isiqu sesicelo (inxalenye eninzi/ifomu-data)
| Iparamitha | Udidi | Ifunekayo | Iinketho ze projekti |
|---|---|---|---|
file |
file | Ewe | Ifayile yesandi/ividiyo emayiguqulwe kancinane |
model |
string | Akukho nanye | Large-v3 (emiselweyo) okanye turbo |
language |
string | Akukho nanye | Ikhowudi yelwimi, okanye ukukhangela ngokuzenzekelayo (okungagqibekanga) |
enable_word_timestamps |
bool | Akukho nanye | Yongeza ixesha lokuqalisa nexesha lokuphelisa kwigama ngalinye |
callback_url |
string | Akukho nanye | I-URL efumana isiphumo nge-POST xa umsebenzi uphele |
I-response
I-id yomsebenzi (uuid). I-Poll GET /v1/transcribe/results/?uuid= de iimeko zigqitywe; impendulo iza kuthatha umbhalo, ulwimi kunye neendawo.
Iimodeli Ezifumanekayo
Iimodeli zombhalo-ukuthetha
| Imodeli ID | Iinketho ze projekti | Elungileyo |
|---|---|---|
kokoro |
I-TTS ekhawulezayo eneelwimi ezininzi | Usebenziso oluqhelekileyo, ulindelwe phantsi |
piper |
Lightweight multilingual TTS (commercially licensed voices only) | Fast renders in 11 languages |
vits |
LJSpeech plus 108 VCTK English speakers | IsiNgesi |
bark |
TTS ejikelezayo, 130 ezimiselweyo kwiilwimi ezili-13 | Uluhlu, unikezelo olubonisayo (olubomvu) |
openvoice |
Ukuphinda usebenzise ilizwi kunye nokuguqulela ilizwi elikhethiweyo | Name=Indlela yokukhetha i-Chatterbox |
chatterbox |
Uhlobo lwesandi olungenalutho olunokulinganisa kwiilwimi ezili-23; imodeli yokuguqulela isandi emiselweyo | Ukuklona ukusuka kwisikhokelo esifutshane |
Iimodeli zokuthetha- ukuya- kumbhalo
| Imodeli ID | Iinketho ze projekti | Elungileyo |
|---|---|---|
whisper-large-v3 |
OpenAI Whisper Large v3 | Igama le projekti |
whisper-turbo |
I-Whisper ekhawulezayo ekhawulezayo | Uguqulelo kolunye ulwimi |
Iikhowudi zemposiso
| Ikhowudi | I-I_status | Iinketho ze projekti |
|---|---|---|
invalid_api_key |
401 | Isitshixo se-API esingasebenziyo okanye esilahlekileyo |
insufficient_credits |
402 | Akukho tyala elininzi lenkqubo |
rate_limit_exceeded |
429 | Iisicelo esininzi kakhulu |
invalid_request |
400 | Isiqu sesicelo esingalunganga |
model_not_found |
404 | Imodeli eceliweyo ayikho |
voice_not_found |
404 | Isithethi esiceliweyo asikho |
file_too_large |
413 | Ifayile elayishwe ngaphezulu idlula umda wobungakanani |
server_error |
500 | Imposiso Yomncedisi Wangaphakathi |