Litokomane tsa API

Ho kenya Speech to Speech AI ho li-application tsa hau ka REST API e bonolo.

Qhobosheane

Ho qala le Speech to Speech AI API ka tlase ho metsotso e 5.

URL ea sehlooho
https://api.speechtospeechai.com/v1
1. Qala mosebetsi

Li-tools tsohle ke POST ka konopo ea hau ea API. Likarabo ke ID ea mosebetsi, eseng audio.

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. Poll bakeng sa phello

E-ba le potso ea mosebetsi ka metsotsoana e seng mekae ho fihlela boemo bo felile, ebe o kenya 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"}

Bopaki

Lipotso tsohle tsa API li hloka ho phetheloa ka ho sebelisa konopo ea API Authorization mohlokomeli.

Mofuta oa _lebokose
Authorization: Bearer YOUR_API_KEY
Fumana konopo ea hau ea API ho tsoa ho dashboard ea ak'haonte.

Ho hloleha ha puisano

POST /v1/voice-clone/

Ho bua ka mongolo ka lentsoe la ho ngola ho khuts'oane.

Bophelo ba kopo (li-multipart/form-data)
Parameter _Tš'oara E hlokahala Bophara:
reference_audio file Ha ho joalo Ho ngola lentsoe ho clone (10-30 sekontiri li khothalletsoa, ho fihlela ho 60)
text string Ha ho joalo Tekanyo ea ho bua ka lentsoe lena
model string Ha ho joalo chatterbox, openvoice, tortoise kapa gpt-sovits
language string Ha ho joalo Kopa ea puo (e tloaelehang: en)
Likarabo

ID ea mosebetsi (uuid). Poll GET /v1/speech/results/?uuid= ho fihlela boemo bo felile'me u romelle result_url.

Text ho To Speech

POST /v1/tts/

Ho fetola tlhaloso ho ho bua ka audio.

Bophelo ba kopo (li-multipart/form-data)
Parameter _Tš'oara E hlokahala Bophara:
text string Ha ho joalo Tebe e tla fetoloa (matsatsi a 5000)
model string Ha ho joalo kokoro (e tloaelehang), piper, vits kapa bark
voice_id string Ha ho joalo ID ea lentsoe ho tloha GET /v1/speech/voices/?model=
language string Ha ho joalo Kopa ea puo (e tloaelehang: en)
speed float Ha ho joalo Mofuta oa lebelo (0.5-2.0, ditlhophiso 1.0)
format string Ha ho joalo wav (e tloaelehileng), mp3, ogg kapa flac. Li-tags tsa SSML li tlosoa pele ho synthesization.
Likarabo

ID ea mosebetsi (uuid). Poll GET /v1/speech/results/?uuid= ho fihlela boemo bo felile'me u romelle result_url.

Ho bua ho ea ho Tekanyo

POST /v1/transcribe/

Ho ngola molumo ho ea ho mongolo ka li-timestamp.

Bophelo ba kopo (li-multipart/form-data)
Parameter _Tš'oara E hlokahala Bophara:
file file Ha ho joalo Faele ea audio/video e tla fetoloa
model string Ha ho joalo large-v3 (e tloaelehileng) kapa turbo
language string Ha ho joalo Kopa ea puo, kapa ho e bona ka ho toba (e tloaelehang)
enable_word_timestamps bool Ha ho joalo Ho eketsa nako ea ho qala le ho fela ho mantsoe ohle
callback_url string Ha ho joalo URL e tla fumana phello ka POST ha mosebetsi o fela
Likarabo

ID ea mosebetsi (uuid). Poll GET /v1/transcribe/results/?uuid= ho fihlela boemo bo felile; karabo e tla ba le mongolo, puo le li-segments.

Li-model tse fumanehang

Li-Models tsa Tebe-ho-Bongata
ID ea Mofuta Bophara: E ntle ka ho fetisisa bakeng sa
kokoro TTS e potlakileng e nang le lipuo tse ngata Ho sebelisa ka ho feletseng, ho se ke ha e-ba le nako e telele
piper TTS e bonolo ea lipuo tse ngata (li-voices tse nang le laesense ea khoebo feela) Ho etsa ka potlako ka lipuo tse 11
vits LJSpeech le 108 VCTK ba buang Senyesemane Li-accent tsa Se-Inglise
bark TTS ea ho theha, li-presets tse 130 ka lipuo tse 13 E fapane, e hlalosang ho fana (ho potlaka)
openvoice Ho hloleha ha puo le ho fetoloa ha puo e sa hlokahaleng E 'ngoe ea likhetho tsa Chatterbox
chatterbox Cloning ea lentsoe la Zero-shot ka lipuo tse 23; moralo oa ho fetolela lentsoe ka ho lekola Ho clone ho tloha ho mohlala o khuts'oane
Li-Models tsa ho bua ho ea ho tokiso
ID ea Mofuta Bophara: E ntle ka ho fetisisa bakeng sa
whisper-large-v3 OpenAI Whisper Large v3 Bophara bo phahameng, lipuo tse ngata
whisper-turbo Whisper e ntlafalitsoeng ho potlaka Ho ngola ka potlako

Li-codes tsa bothata

Li-code _Tlhahlobo Bophara:
invalid_api_key 401 Setsi sa API se sa nepahale kapa se sa fumaneha
insufficient_credits 402 Ha ho credits tse lekaneng bakeng sa ts'ebetso ena
rate_limit_exceeded 429 Lipotso tse ngata haholo
invalid_request 400 Bophelo ba kopo bo sa nepahale
model_not_found 404 Mofuta o hlokahaleng ha o na
voice_not_found 404 Sehlooho se boletsoeng se sa na le
file_too_large 413 Faele e kentsoeng e feta boholo ba eona
server_error 500 Phapang ea server ea kahare