curl https://api.siati.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer $SIATI_API_KEY" \
  -F file=@messaggio.m4a \
  -F model=whisper-1 \
  -F language=it

from openai import OpenAI

client = OpenAI(
    base_url="https://api.siati.ai/v1",
    api_key="sk-siati-...",
)

with open("messaggio.m4a", "rb") as f:
    result = client.audio.transcriptions.create(
        model="whisper-1",
        file=f,
        language="it",                  # opzionale, auto-detect altrimenti
        response_format="json",         # json | verbose_json | text
        prompt="vocabolario tecnico",   # hint per migliorare trascrizione
    )

print(result.text)

const form = new FormData();
form.append('file', audioBlob, 'messaggio.m4a');
form.append('model', 'whisper-1');
form.append('language', 'it');

const res = await fetch('https://api.siati.ai/v1/audio/transcriptions', {
  method: 'POST',
  headers: { 'Authorization': `Bearer ${SIATI_API_KEY}` },
  body: form,
});
const { text } = await res.json();
console.log(text);

let url = URL(string: "https://api.siati.ai/v1/audio/transcriptions")!
let boundary = UUID().uuidString
var req = URLRequest(url: url)
req.httpMethod = "POST"
req.setValue("Bearer \(apiKey)", forHTTPHeaderField: "Authorization")
req.setValue("multipart/form-data; boundary=\(boundary)", forHTTPHeaderField: "Content-Type")

var body = Data()
// audio file part
body.append("--\(boundary)\r\n".data(using: .utf8)!)
body.append("Content-Disposition: form-data; name=\"file\"; filename=\"audio.m4a\"\r\n".data(using: .utf8)!)
body.append("Content-Type: audio/m4a\r\n\r\n".data(using: .utf8)!)
body.append(audioData)
body.append("\r\n".data(using: .utf8)!)
// model + language fields
for (k, v) in [("model", "whisper-1"), ("language", "it")] {
    body.append("--\(boundary)\r\n".data(using: .utf8)!)
    body.append("Content-Disposition: form-data; name=\"\(k)\"\r\n\r\n\(v)\r\n".data(using: .utf8)!)
}
body.append("--\(boundary)--\r\n".data(using: .utf8)!)
req.httpBody = body

let (data, _) = try await URLSession.shared.data(for: req)
let result = try JSONDecoder().decode(TranscriptionResponse.self, from: data)
print(result.text)

{
  "text": "Ciao, come stai oggi? Volevo chiederti se possiamo vederci domani."
}

{
  "task": "transcribe",
  "language": "it",
  "duration": 4.32,
  "text": "Ciao, come stai oggi? Volevo chiederti se possiamo vederci domani.",
  "segments": [
    {
      "id": 0,
      "start": 0.0,
      "end": 2.1,
      "text": "Ciao, come stai oggi?",
      "avg_logprob": -0.21,
      "no_speech_prob": 0.02
    },
    {
      "id": 1,
      "start": 2.1,
      "end": 4.32,
      "text": "Volevo chiederti se possiamo vederci domani."
    }
  ]
}

curl https://api.siati.ai/v1/audio/speech \
  -H "Authorization: Bearer $SIATI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tts-1",
    "voice": "paola",
    "input": "Ciao! Sono Paola, la voce italiana di SIATI.",
    "response_format": "wav"
  }' \
  --output benvenuto.wav

res = client.audio.speech.create(
    model="tts-1",                        # accettato per compat OpenAI
    voice="paola",                        # paola, riccardo, amy, ryan, thorsten, siwis
    input="Ciao! Sono Paola.",
    response_format="wav",                # wav | mp3
    # speed=1.0,                          # accettato ma ignorato dal MVP
)

res.stream_to_file("benvenuto.wav")

const res = await fetch('https://api.siati.ai/v1/audio/speech', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${SIATI_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
    model: 'tts-1',
    voice: 'paola',
    input: 'Ciao!',
    response_format: 'mp3',
  }),
});
const audioBlob = await res.blob();
// In browser:
const url = URL.createObjectURL(audioBlob);
new Audio(url).play();

Content-Type: audio/wav     (o audio/mpeg per mp3)
Content-Length: 152043
X-Request-Id: req_abc123...

curl https://my.siati.ai/api/v1/audio/transcribe \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -F audio=@nota.m4a \
  -F language=it \
  -F with_segments=true

{
  "text": "Promemoria: chiamare Marco alle 15.",
  "language": "it",
  "duration_seconds": 3.4,
  "model": "whisper-1",
  "segments": [...]
}

curl https://my.siati.ai/api/v1/audio/synthesize \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"text":"Ciao!","voice":"paola","format":"mp3"}' \
  --output out.mp3

curl https://my.siati.ai/api/v1/audio/voices \
  -H "Authorization: Bearer $JWT_TOKEN"

{
  "voices": [
    { "name": "paola", "language": "it", "gender": "female" },
    { "name": "riccardo", "language": "it", "gender": "male" },
    ...
  ],
  "defaults_by_language": {
    "it": "paola",
    "en": "amy",
    "de": "thorsten",
    "fr": "siwis",
    "rm": "paola"
  }
}

# 1. Trascrivi voice message utente
with open("user-message.m4a", "rb") as f:
    transcript = client.audio.transcriptions.create(model="whisper-1", file=f, language="it")

# 2. Chiedi al LLM
resp = client.chat.completions.create(
    model="apertus-70b-instruct",
    messages=[{"role": "user", "content": transcript.text}],
)
ai_text = resp.choices[0].message.content

# 3. Sintetizza la risposta
audio = client.audio.speech.create(model="tts-1", voice="paola", input=ai_text)
audio.stream_to_file("ai-reply.mp3")

client.audio.transcriptions.create(
    model="whisper-1-hd",
    file=open("riunione.m4a", "rb"),
    language="it",
    prompt="SIATI, Apertus, Mistral, Qdrant, GPU, embedding, RAG, tier",
    response_format="verbose_json",
)