音声を複製する

import base64

from google import genai

client = genai.Client(enterprise=True, project="PROJECT_ID", location="global")

with open("SOURCE_AUDIO_PATH", "rb") as f:
    source_b64 = base64.b64encode(f.read()).decode("utf-8")

with open("CONSENT_AUDIO_PATH", "rb") as f:
    consent_b64 = base64.b64encode(f.read()).decode("utf-8")

replicated_voice = client.voices.create(
    store=True,
    voice={
        "type": "VOICE_TYPE_REPLICATED",
        "display_name": "Custom replicated speaker",
        "replicated": {
            "source_audio": {"mime_type": "audio/wav", "data": source_b64},
            "consent_audio": {"mime_type": "audio/wav", "data": consent_b64},
        },
    },
    timeout=60,
)

# The voice ID starts with "voice_".
print(f"Created voice ID: {replicated_voice.id}")

SOURCE_B64=$(base64 -w 0 SOURCE_AUDIO_PATH)
CONSENT_B64=$(base64 -w 0 CONSENT_AUDIO_PATH)

cat > request.json <<EOF
{
  "store": true,
  "voice": {
    "type": "VOICE_TYPE_REPLICATED",
    "displayName": "Custom replicated speaker",
    "replicated": {
      "sourceAudio": {"mimeType": "audio/wav", "data": "${SOURCE_B64}"},
      "consentAudio": {"mimeType": "audio/wav", "data": "${CONSENT_B64}"}
    }
  }
}
EOF

curl -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  https://aiplatform.googleapis.com/v1beta1/projects/PROJECT_ID/locations/global/voices \
  -d @request.json | jq -r '.id'

import base64

from google import genai

client = genai.Client(enterprise=True, project="PROJECT_ID", location="global")

with open("SOURCE_AUDIO_PATH", "rb") as f:
    source_b64 = base64.b64encode(f.read()).decode("utf-8")

with open("CONSENT_AUDIO_PATH", "rb") as f:
    consent_b64 = base64.b64encode(f.read()).decode("utf-8")

replicated_voice = client.voices.create(
    store=False,
    voice={
        "model": "gemini-3.8-flash-tts",
        "type": "VOICE_TYPE_REPLICATED",
        "replicated": {
            "source_audio": {"mime_type": "audio/wav", "data": source_b64},
            "consent_audio": {"mime_type": "audio/wav", "data": consent_b64},
        },
    },
    timeout=60,
)

# Store the key in your application. It starts with "voicekey_".
voice_key = replicated_voice.key

SOURCE_B64=$(base64 -w 0 SOURCE_AUDIO_PATH)
CONSENT_B64=$(base64 -w 0 CONSENT_AUDIO_PATH)

cat > request.json <<EOF
{
  "store": false,
  "voice": {
    "model": "gemini-3.8-flash-tts",
    "type": "VOICE_TYPE_REPLICATED",
    "replicated": {
      "sourceAudio": {"mimeType": "audio/wav", "data": "${SOURCE_B64}"},
      "consentAudio": {"mimeType": "audio/wav", "data": "${CONSENT_B64}"}
    }
  }
}
EOF

curl -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  https://aiplatform.googleapis.com/v1beta1/projects/PROJECT_ID/locations/global/voices \
  -d @request.json | jq -r '.key'

from google import genai

client = genai.Client(enterprise=True, project="PROJECT_ID", location="global")

response = client.models.generate_content(
    model="gemini-3.8-flash-tts",
    contents=[{
        "role": "user",
        "parts": [{
            "text": "Hello! This audio was generated with a replicated voice.",
            "speech_metadata": {"style": "warm and conversational"},
        }],
    }],
    config={
        "response_modalities": ["AUDIO"],
        "speech_config": {"voice_config": {"voice": "VOICE"}},
    },
)

# The SDK has already decoded the base64 audio, so inline_data.data is a
# complete WAV file by default.
with open("replicated_voice.wav", "wb") as f:
    f.write(response.candidates[0].content.parts[0].inline_data.data)

curl -X POST \
  -H "Authorization: Bearer $(gcloud auth print-access-token)" \
  -H "Content-Type: application/json" \
  https://aiplatform.googleapis.com/v1/projects/PROJECT_ID/locations/global/publishers/google/models/gemini-3.8-flash-tts:generateContent \
  -d '{
    "contents": [{
      "role": "user",
      "parts": [{
        "text": "Hello! This audio was generated with a replicated voice.",
        "speechMetadata": {"style": "warm and conversational"}
      }]
    }],
    "generationConfig": {
      "responseModalities": ["AUDIO"],
      "speechConfig": {
        "voiceConfig": {"voice": "VOICE"}
      }
    }
  }' | jq -r '.candidates[0].content.parts[0].inlineData.data' | base64 --decode > replicated_voice.wav