Skip to main content

Overview

Voice models determine how your AI agents sound during phone calls. Call24x7.AI supports multiple voice providers and a wide variety of voices to match your brand and use case.

Available Voice Providers

Call24x7.AI integrates with leading voice synthesis providers:
  • Cartesia: High-quality, natural-sounding voices
  • ElevenLabs: Premium voice cloning and synthesis
  • DeepDub: Multilingual voice solutions

Selecting a Voice

Voice Characteristics

Consider these factors when choosing a voice:
  • Gender: Male, female, or neutral
  • Age: Young, middle-aged, or mature
  • Tone: Professional, friendly, calm, energetic
  • Accent: Regional accents and pronunciations
  • Language: Native language support

Matching Voice to Use Case

Customer Service

Use warm, professional voices that convey empathy and helpfulness

Sales & Marketing

Choose energetic, engaging voices that capture attention

Technical Support

Select clear, articulate voices that are easy to understand

Healthcare

Use calm, reassuring voices that build trust

Configuring Voice in Agents

Using the Developer Portal

  1. Navigate to Agents
  2. Select or create an agent
  3. Go to Voice Settings
  4. Choose your voice provider
  5. Select a specific voice
  6. Preview the voice
  7. Save your selection

Using the API

When creating agents via API, specify the voice configuration in the agent settings.

Voice Customization

Speaking Rate

Adjust how fast or slow the agent speaks:
  • Normal: Standard speaking rate
  • Slow: Slower for clarity or emphasis
  • Fast: Faster for efficiency (use carefully)

Emphasis and Intonation

Some voice providers support:
  • Emphasis markers: Add emphasis to specific words
  • Pauses: Control pauses for natural flow
  • Intonation: Adjust tone for questions vs. statements

Language Support

Voice models support multiple languages:
  • English: Multiple accents (US, UK, Australian, etc.)
  • Spanish: Various regional variations
  • French: European and Canadian French
  • German: Standard and regional German
  • And more: Check provider documentation for full language list

Multilingual Agents

Configure agents to:
  • Detect the caller’s language
  • Switch languages mid-conversation
  • Maintain consistent voice across languages

Voice Quality Considerations

Audio Quality

  • Sample Rate: Higher sample rates provide better quality
  • Bitrate: Higher bitrates for clearer audio
  • Latency: Balance between quality and real-time performance

Naturalness

Modern voice synthesis provides:
  • Natural pauses: Realistic breathing and pauses
  • Emotional expression: Convey emotions appropriately
  • Pronunciation: Accurate pronunciation of names and terms

Testing Voices

Before deploying an agent:
  1. Preview: Listen to voice samples in the portal
  2. Test Calls: Make test calls with different voices
  3. Get Feedback: Gather feedback from team members
  4. Compare: Test multiple voices for the same use case

Voice Provider Details

Cartesia

  • Strengths: High quality, low latency, natural-sounding
  • Best for: Real-time conversations, customer service
  • Languages: Multiple languages supported

ElevenLabs

  • Strengths: Voice cloning, premium quality, emotional range
  • Best for: Brand-specific voices, high-end applications
  • Languages: Extensive language support

DeepDub

  • Strengths: Multilingual, accent support, regional variations
  • Best for: International operations, diverse customer base
  • Languages: Wide range of languages and accents

Best Practices

  1. Match Brand: Choose voices that align with your brand identity
  2. Test Thoroughly: Test voices with real call scenarios
  3. Consider Audience: Select voices your audience will find relatable
  4. Consistency: Use consistent voices across similar use cases
  5. Accessibility: Ensure voices are clear and easy to understand
  6. Update Regularly: Review and update voice selections periodically

Troubleshooting

  • Try a different voice from the same provider
  • Adjust speaking rate
  • Review agent instructions for better flow
  • Consider switching voice providers
  • Verify language settings match the voice
  • Check that the voice supports the target language
  • Test with a different voice for that language
  • Check network connection
  • Verify voice provider settings
  • Test with different voices
  • Contact support if issues persist

Voice Samples

Listen to voice samples in the Developer Portal to help you choose the right voice for your use case.

Support

For voice-related questions: