API Dokumentasie

Geldigheidsverklaring

Vir premiese modelle en hoër tempogrense, met â n Beer-bewys in die Authorization kop.

Authorization: Bearer sk-tts-your-api-key-here

SDKs

Official SDKs make it easy to integrate TTS.ai into your application. Both are open source and available on GitHub.

pip install ttsai

from tts_ai import TTSClient

client = TTSClient(api_key="sk-tts-...")
audio = client.generate(
    text="Hello world!",
    model="kokoro"
)
client.save(audio, "output.wav")

npm install @ttsainpm/ttsai

const { TTSClient } = require('@ttsainpm/ttsai');

const client = new TTSClient({
  apiKey: 'sk-tts-...'
});
const audio = await client.generate({
  input: 'Hello world!',
  model: 'kokoro'
});
await client.saveToFile(audio, 'output.wav');

Basis URL

Alle eindpunte is relatief tot hierdie basis URL. Byvoorbeeld, die TTS-punt is:

Tempo Limiet

API tempo perke wissel deur plan:

Herhaalde beperk opskrifte word ingesluit in elke antwoord: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset.

Kredietkoste

Teks vir spraak

Skakel teks om na toespraak oudio. Gee terug oudio-lêer in die versoekte formaat.

Versoek Liggaam

Voorbeeld Versoek

curl -X POST https://api.tts.ai/v1/tts/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "kokoro",
    "text": "Hello from TTS.ai! This is a test.",
    "voice": "af_bella",
    "format": "mp3"
  }' \
  --output output.mp3

SSML etiket

Oorvloei nommers, datums, geld eenheid, telefoonnommers en akronieme in

Datum formaat verstek na mid vir Engels en elders; oorskryf met format=\

{
  "model": "kokoro",
  "voice": "af_bella",
  "text": "Your appointment is on <say-as interpret-as=\"date\">2026-04-26</say-as> at <say-as interpret-as=\"time\">14:30</say-as>. Please call <say-as interpret-as=\"telephone\">+1-555-867-5309</say-as> if you need to reschedule."
}

Reaksie

The TTS endpoint queues your request and returns a JSON response with a job UUID. You then poll for the result.

Step 1: Submit request

{
  "uuid": "77b71db532874ce98e84a69a2d740d4c",
  "job_id": "f21316bb-aefa-480d-8523-701d1e3184ce",
  "status": "queued",
  "credits_used": 11,
  "credits_remaining": 15000
}

Step 2: Poll for result

Poll this endpoint every 1-2 seconds until status is completed or failed.

{
  "status": "completed",
  "result_url": "https://api.tts.ai/static/downloads/77b71db5.../output.mp3"
}

{
  "status": "processing"
}

Step 3: Download audio

Fetch the result_url from the completed response to download the audio file.

Volgrote voorbeeld

import requests, time

API_KEY = "sk-tts-your-key"
BASE = "https://api.tts.ai"

# 1. Submit TTS request
resp = requests.post(f"{BASE}/v1/tts/", json={
    "model": "kokoro",
    "text": "Hello from TTS.ai!",
    "voice": "af_bella"
}, headers={"Authorization": f"Bearer {API_KEY}"})
data = resp.json()
uuid = data["uuid"]

# 2. Poll for result
while True:
    result = requests.get(f"{BASE}/v1/speech/results/",
        params={"uuid": uuid}).json()
    if result["status"] == "completed":
        # 3. Download audio
        audio = requests.get(result["result_url"])
        with open("output.mp3", "wb") as f:
            f.write(audio.content)
        break
    elif result["status"] == "failed":
        raise Exception(result.get("error", "Generation failed"))
    time.sleep(1.5)

Streaming alternative: For supported models (Kokoro, MeloTTS), use POST /v1/tts/stream/ for real-time Server-Sent Events (SSE) streaming — no polling needed.

Woorde tot teks

Ingeskryf klank na teks. Ondersteun 99 tale met outo- opspoor.

Versoek Liggaam (multipart/form-data)

Reaksie

{
  "text": "Hello, this is a transcription test.",
  "language": "en",
  "duration": 3.5,
  "segments": [
    {
      "start": 0.0,
      "end": 1.8,
      "text": "Hello, this is",
      "speaker": "SPEAKER_00"
    },
    {
      "start": 1.8,
      "end": 3.5,
      "text": "a transcription test.",
      "speaker": "SPEAKER_00"
    }
  ]
}

Stemverkleuring

Genereer spraak in 'n gekloonde stem. Oplaai' n verwysing oudio en teks.

Versoek Liggaam (multipart/form-data)

Reaksie

Gee terug die oudiolêer as binêre data, dieselfde as die TTS-punt.

Stemveranderder

Skakel oudio om na klank soos 'n ander stem. Oplaai bron-oudio en kies 'n teikenstem.

Versoek Liggaam (multipart/form-data)

Voorbeeld Versoek

curl -X POST https://api.tts.ai/v1/voice-convert/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -F "file=@source_audio.mp3" \
  -F "target_voice=af_bella" \
  -F "model=openvoice" \
  -o converted.wav

Reaksie

Gee terug die omgeskakelde oudiolêer as binêre data.

Spraak - vertaling

Vertaal die klank van een taal na 'n ander. Kombineer spraak-na-teks, vertaling en teks-na-sech in' n enkele roep.

Versoek Liggaam (multipart/form-data)

Reaksie

{
  "original_text": "Hello, how are you?",
  "translated_text": "Hola, como estas?",
  "source_language": "en",
  "target_language": "es",
  "audio_url": "https://api.tts.ai/v1/results/translate_abc123.mp3",
  "credits_used": 5
}

Spraak tot spraak

Verander spraakstyl, emosie of aflewering terwyl jy die inhoud behou.'n Nuttige manier om'n mens se stemtoon, stemtoon en gevoel te verander.

Versoek Liggaam (multipart/form-data)

Reaksie

Gee terug die veranderde oudiolêer as binêre data.

Oudiogereedskap

Oudioverwerkingspunte vir verbeteringe, sangverwydering, stamverdeeling en nog meer.

{
  "key": "C",
  "scale": "Major",
  "bpm": 120.0,
  "time_signature": "4/4",
  "camelot": "8B",
  "compatible_keys": ["C Major", "G Major", "F Major", "A Minor"]
}

Stem Gesels

Stuur oudio- of teks en ontvang 'n kunsmatige antwoord met sintetiseerde spraak.

Versoek Liggaam (multipart/form-data of JSON)

Reaksie

{
  "conversation_id": "conv_abc123",
  "user_text": "What is the capital of France?",
  "ai_text": "The capital of France is Paris.",
  "audio_url": "https://api.tts.ai/v1/audio/tmp/resp_xyz.mp3",
  "credits_used": 3
}

BangTTS

Dien veelvuldige tekste vir parallelle TTS - geslag in. Ontvang op'n mooi manier'n webhoek - terugroep wanneer alle werke voltooi is.

Parameters

Reaksie

{
  "batch_id": "abc123",
  "total": 3,
  "completed": 0,
  "status": "processing"
}

Posl vordering met KRY /v1/ts/bat/result/?bang_id=abc123

Stem inbed

Pre- compute 'n stem vergieting van verwysing oudio. Gebruik die teruggekeerde embed_id in latere stem kloning versoeke vir naby-instant geslag.

Parameters

Reaksie

{
  "embed_id": "emb_abc123",
  "model": "chatterbox",
  "duration_ms": 450
}

Gesondheidstoets

Bevestig GPU bediener status, gelaaide modelle, en wagtou grootte. Nee geldigheidsverklaring benodig. Kasd vir 30 sekondes.

Reaksie

{
  "status": "online",
  "latency_ms": 45,
  "queue_size": 3,
  "models_loaded": ["kokoro", "chatterbox", "cosyvoice2"]
}

Lysmodel's

Gee terug 'n lys van alle beskikbare modelle met hul vermoëns.

Reaksie

{
  "models": [
    {
      "id": "kokoro",
      "name": "Kokoro",
      "type": "tts",
      "tier": "standard",
      "languages": ["en", "ja", "ko", "zh", "fr"],
      "supports_cloning": false,
      "supports_streaming": true,
      "credits_per_1k_chars": 2
    },
    {
      "id": "chatterbox",
      "name": "Chatterbox",
      "type": "tts",
      "tier": "premium",
      "languages": ["en"],
      "supports_cloning": true,
      "supports_streaming": true,
      "credits_per_1k_chars": 4
    }
  ]
}

Maak 'n mens se stem oop

Gee terug 'n lys van alle beskikbare stemme, verkies gegefilter deur model of taal.

Navraag Parameters

Reaksie

{
  "voices": [
    {
      "id": "af_bella",
      "name": "Bella",
      "model": "kokoro",
      "language": "en",
      "gender": "female",
      "preview_url": "https://api.tts.ai/v1/voices/preview/af_bella.mp3"
    }
  ],
  "total": 142
}

Subtitels (RT / VTT) nuwe

Genereer geroniseerde subtitels vir enige voltooide TTS werk. Laat loop Whiper belyning oor die oudio en gee terug SrT of WebVT. Resultaat is kasgeheue op disket sodat 'n tweede roep vir dieselfde uid is' n disket lees.

Navraag Parameters

curl "https://api.tts.ai/v1/speech/subtitles/?uuid=$UUID&format=srt&download=1" -o subtitles.srt

Woordeboeke nuwe

Sê vir die TTS-enjin hoe om spesifieke woorde uit te spreek. Gestoor inskrywings outo- apply na elke TTS versoek jy maak. 200-entry per- per-telling beperk.

Versoek Liggaam (POST)

# Save an entry
curl -X POST https://tts.ai/api/v1/pronunciations/ \
  -H "Authorization: Bearer sk-tts-..." \
  -H "Content-Type: application/json" \
  -d '{"word": "GIF", "replacement": "jiff"}'

# List your entries
curl https://tts.ai/api/v1/pronunciations/ -H "Authorization: Bearer sk-tts-..."

# Delete entry by id
curl -X DELETE "https://tts.ai/api/v1/pronunciations/?id=42" -H "Authorization: Bearer sk-tts-..."

Jy kan ook verby per-request oorskryf sonder om hulle te red ${ sluit Profices op enige /v1/ts/ roep as 'n voorwerp of 'n skikking (sien die TTS- eindig params) in.

Artikel Narator nuwe

Laat val 'n enkel < etiket op enige artikel bladsy en besoekers kry 'n vaste leser balk wat die bladsy op kliek beskryf. Outo-speur die artikel liggaam, ondersteun pasgemaakte stem / model / posisie / aksent kleur.

<script src="https://tts.ai/narrator.js"
    data-pk="pk-tts-your-publishable-key"
    data-voice="af_bella"
    data-model="kokoro"
    data-extract="auto"
    data-position="bottom"
    data-color="#e60000"
    data-locale="en"></script>

Keuses

Bron op GitHub:

• Python
• JavaScript
• cURL

import requests

API_KEY = "sk-tts-your-key"

# Text to Speech
response = requests.post(
    "https://api.tts.ai/v1/tts/",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "model": "kokoro",
        "text": "Hello from TTS.ai!",
        "voice": "af_bella",
        "format": "mp3"
    }
)

with open("output.mp3", "wb") as f:
    f.write(response.content)

print(f"Credits used: {response.headers.get('X-Credits-Used')}")

# Speech to Text
with open("recording.mp3", "rb") as f:
    response = requests.post(
        "https://api.tts.ai/v1/stt/",
        headers={"Authorization": f"Bearer {API_KEY}"},
        files={"file": f},
        data={"model": "faster-whisper", "timestamps": "true"}
    )

result = response.json()
print(result["text"])

# Voice Cloning
with open("reference.wav", "rb") as ref:
    response = requests.post(
        "https://api.tts.ai/v1/tts/clone/",
        headers={"Authorization": f"Bearer {API_KEY}"},
        files={"reference_audio": ref},
        data={
            "text": "This speech uses a cloned voice.",
            "model": "chatterbox"
        }
    )

with open("cloned_output.mp3", "wb") as f:
    f.write(response.content)

const API_KEY = 'sk-tts-your-key';

// Text to Speech
const response = await fetch('https://api.tts.ai/v1/tts/', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${API_KEY}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    model: 'kokoro',
    text: 'Hello from TTS.ai!',
    voice: 'af_bella',
    format: 'mp3'
  })
});

const audioBlob = await response.blob();
const audioUrl = URL.createObjectURL(audioBlob);
const audio = new Audio(audioUrl);
audio.play();

// Speech to Text
const formData = new FormData();
formData.append('file', audioFile);
formData.append('model', 'faster-whisper');

const response = await fetch('https://api.tts.ai/v1/stt/', {
  method: 'POST',
  headers: { 'Authorization': `Bearer ${API_KEY}` },
  body: formData
});

const result = await response.json();
console.log(result.text);

# Text to Speech
curl -X POST https://api.tts.ai/v1/tts/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -H "Content-Type: application/json" \
  -d '{"model":"kokoro","text":"Hello!","voice":"af_bella","format":"mp3"}' \
  -o output.mp3

# Speech to Text
curl -X POST https://api.tts.ai/v1/stt/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -F "file=@recording.mp3" \
  -F "model=faster-whisper" \
  -F "timestamps=true"

# Voice Cloning
curl -X POST https://api.tts.ai/v1/tts/clone/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -F "reference_audio=@reference.wav" \
  -F "text=This uses a cloned voice." \
  -F "model=chatterbox" \
  -o cloned.mp3

# Audio Enhancement
curl -X POST https://api.tts.ai/v1/audio/enhance/ \
  -H "Authorization: Bearer sk-tts-your-key" \
  -F "file=@noisy_audio.mp3" \
  -F "denoise=true" \
  -F "enhance_clarity=true" \
  -o enhanced.mp3

Foutkodes

Alle foute gee terug 'n JSON antwoord met' n error veld.

{
  "error": {
    "code": "insufficient_credits",
    "message": "You do not have enough characters for this request.",
    "characters_required": 4000,
    "characters_available": 2000
  }
}

Uitpakkies

Vir lang-ring take (verdeelde, klomp TTS) kan jy verskaf 'n webhoek_ url parameter. Wanneer die opdrag voltooi, sal ons die resultaat na jou URL toe stuur.

{
  "event": "task.completed",
  "task_id": "task_abc123",
  "status": "success",
  "result_url": "https://api.tts.ai/v1/results/task_abc123",
  "credits_used": 12,
  "created_at": "2025-01-15T10:30:00Z",
  "completed_at": "2025-01-15T10:30:45Z"
}