Speech synthesis API

GET /api/v1/voices

List available TTS voices (id, language, engine metadata).

GET /api/v1/languages

Languages aligned with the translate + voice catalog.

POST /api/v1/tts

Synthesize speech. Body JSON:

{
  "text": "Hello from Audiozora",
  "voice": "af_heart",
  "rate": "+0%",
  "pitch": "+0Hz"
}

Response includes url (WAV under /workspace/uploads/…), chars, and source.

curl -X POST https://audiozora.com/api/v1/tts \
  -H "Authorization: Bearer az_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text":"Natural text to speech from the API.","voice":"af_heart"}'

POST /api/v1/translate

Translate text before TTS (multilingual pipeline).

{
  "text": "Bonjour le monde",
  "target_lang": "en-US",
  "source_lang": null
}

GET /api/v1/voice-clone/limits

Sample length guidance, supported formats/languages, and max text length for XTTS-v2 instant clone.

POST /api/v1/voice-clone

Multipart upload: reference sample + text → cloned WAV. Use a clear 3–30s sample (mp3/wav/m4a). Distinct from studio Voice Replace (timing lock on existing media).

curl -X POST https://audiozora.com/api/v1/voice-clone \
  -H "Authorization: Bearer az_YOUR_KEY" \
  -F "text=Hello from my cloned voice" \
  -F "language=en" \
  -F "sample=@my_voice_sample.wav"

POST /api/v1/tts-with-clone

JSON variant with sample_base64 for automation (same engine as /voice-clone).

{
  "text": "Hello from my cloned voice",
  "language": "en",
  "filename": "sample.wav",
  "sample_base64": "<base64 audio bytes>"
}