Voice Mapping

GateFlow provides standardized voice personas that automatically map to the best equivalent voice on each TTS provider.

Standard Voices

Voice ID	Persona	Characteristics
professional	Business	Clear, authoritative, confident
friendly	Conversational	Warm, approachable, welcoming
calm	Soothing	Measured, relaxed, reassuring
energetic	Upbeat	Enthusiastic, dynamic, engaging
serious	Formal	Deliberate, precise, professional
casual	Relaxed	Natural, informal, easy-going

Provider Mappings

OpenAI TTS

GateFlow Voice	OpenAI Voice
professional	onyx
friendly	alloy
calm	nova
energetic	shimmer
serious	fable
casual	echo

ElevenLabs

GateFlow Voice	ElevenLabs Voice
professional	josh
friendly	rachel
calm	bella
energetic	antoni
serious	adam
casual	domi

Using Standard Voices

Basic Usage

response = client.audio.speech.create(
    model="tts-1",
    input="Hello, welcome to our service.",
    voice="friendly"  # Maps to provider-specific voice
)

With Provider Override

response = client.audio.speech.create(
    model="eleven_multilingual_v2",
    input="Hello, welcome to our service.",
    voice="friendly"  # Maps to "rachel" on ElevenLabs
)

Direct Voice ID

Use provider-specific voice IDs directly:

# ElevenLabs voice ID
response = client.audio.speech.create(
    model="eleven_multilingual_v2",
    input="Hello!",
    voice="21m00Tcm4TlvDq8ikWAM"  # Rachel's voice ID
)

# OpenAI voice name
response = client.audio.speech.create(
    model="tts-1",
    input="Hello!",
    voice="alloy"
)

Custom Voice Mappings

Create your own voice mappings:

curl -X POST https://api.gateflow.ai/v1/management/voice-mappings \
  -H "Authorization: Bearer gw_prod_admin_key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "customer-support",
    "mappings": {
      "openai": "nova",
      "elevenlabs": "21m00Tcm4TlvDq8ikWAM"
    }
  }'

Use your custom mapping:

response = client.audio.speech.create(
    model="auto",
    input="How can I help you today?",
    voice="customer-support"
)

Voice Cloning (ElevenLabs)

Use cloned voices with ElevenLabs:

response = client.audio.speech.create(
    model="eleven_multilingual_v2",
    input="Hello from your custom voice!",
    voice="your_cloned_voice_id"
)

Voice Settings

Adjust voice parameters:

response = client.audio.speech.create(
    model="eleven_multilingual_v2",
    input="Hello!",
    voice="friendly",
    extra_body={
        "gateflow": {
            "voice_settings": {
                "stability": 0.5,
                "similarity_boost": 0.75,
                "style": 0.5,
                "speed": 1.0
            }
        }
    }
)

Setting	Range	Description
stability	0.0-1.0	Higher = more consistent, lower = more expressive
similarity_boost	0.0-1.0	How closely to match original voice
style	0.0-1.0	Exaggerate voice style (ElevenLabs only)
speed	0.5-2.0	Speech rate multiplier

Multi-Language Support

Standard voices support multiple languages:

# Same voice, different language
response = client.audio.speech.create(
    model="eleven_multilingual_v2",
    input="Bonjour, comment puis-je vous aider?",
    voice="friendly",
    extra_body={
        "gateflow": {
            "language": "fr"
        }
    }
)

Next Steps

• Pipeline Templates - Pre-configured voice pipelines
• Streaming Speech - Real-time audio generation