Semantic Turn Detection

Semantic turn detection handles cases where users speak incomplete sentences, allowing for more natural conversation flow by waiting for complete thoughts before processing.

Usage

const agent = new OpenaiAgent({
  apiKey: process.env.OPENAI_API_KEY || '',
  systemPrompt: 'You are a helpful assistant',
  autoSemanticTurn: true, // Wait for complete thoughts
  // or provide custom prompt:
  // autoSemanticTurn: 'Last user message is an incomplete sentence',
})

Configuration Options

Option	Type	Description
autoSemanticTurn	boolean | string	Enable detection or provide custom prompt

• true: Uses the default incomplete sentence detection logic
• string: Provides a custom prompt for the AI to use when determining if a message is incomplete

How it Works

The semantic turn detection feature analyzes user messages to determine if they represent complete thoughts or incomplete sentences. When an incomplete sentence is detected, the agent emits a SkipAnswer event to skip the answer.

Example Scenarios

Without Semantic Turn Detection:

• User: "I want to..."
• Agent: "What would you like to do?"
• User: "book a flight"

With Semantic Turn Detection:

• User: "I want to..."
• Agent waits
• User: "book a flight"
• Agent: "I'd be happy to help you book a flight. Where would you like to go?"

Best Practices

• Use this feature for conversational applications where users might pause mid-sentence
• Consider the speech-to-text accuracy when enabling this feature
• Test with various incomplete sentence patterns to ensure proper detection
• Balance responsiveness with waiting for complete thoughts