Skip to main content

Setting Up Voice Simulations

To get started with voice simulations in Maxim:
  • Add a voice provider: Navigate to Settings > Voice Providers and configure your preferred provider (Twilio or VAPI) with the necessary account credentials for secure connection
  • Configure your voice agent: Go to Agents > Voice Agents and create a new agent with its name, description, phone number, and country code
  • Define test scenarios: Create custom scenarios that specify the conversation context, goals, and desired persona characteristics including voice style and behavior patterns
  • Set conversation flow: Configure whether your simulation agent or voice agent initiates the conversation (default is simulation agent waiting for your voice agent to speak first)

Evaluation Capabilities

Maxim’s voice simulation runs provide comprehensive insights into your agent’s performance:
  • Real-time audio streaming and transcription for both simulation and voice agents
  • Complete conversation transcripts displayed side-by-side for easy analysis
  • Audio recordings available for playback and download
  • Statistical conversation metrics including average response latency, average pitch, talk ratio, and speech rate
  • Built-in voice evaluators to assess conversation quality across scenarios
This allows you to identify issues with response timing, conversation flow, and voice quality before deploying to production, ensuring your voice agent delivers consistent, high-quality interactions across diverse user scenarios.