Noise Filtering

User noise filtering automatically filters out meaningless sounds like “uh”, “hmm”, “ahem” that don’t carry conversational meaning, improving the quality of conversation flow.

Usage

const agent = new OpenaiAgent({
  apiKey: process.env.OPENAI_API_KEY || '',
  systemPrompt: 'You are a helpful assistant',
  autoIgnoreUserNoise: true, // Ignore filler sounds
  // or provide custom prompt:
  // autoIgnoreUserNoise: 'Last user message is just an interjection',
})

Configuration Options

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

How it Works

A specific tool is added to the query, that will be called when the agent detects that the user message is just noise. When a noise is detected, the agent automatically emits an CancelLastUserMessage to removed and ignore the last user message.

Common Filtered Sounds

The system typically filters out:

• Filler words: “uh”, “um”, “er”
• Interjections: “hmm”, “ahem”, “oh”
• Breathing sounds and background noise
• Partial words or unclear utterances

Example Scenarios

Without User Noise Filtering:

• User: “Um…”
• Agent: “I’m sorry, I didn’t understand. Could you please repeat that?”
• User: “I want to book a flight”

With User Noise Filtering:

• User: “Um…”
• Agent ignores the noise
• User: “I want to book a flight”
• Agent: “I’d be happy to help you book a flight. Where would you like to go?”

Best Practices

• Consider the quality of your speech-to-text service when configuring
• Test with various noise patterns to ensure legitimate short responses aren’t filtered
• Balance filtering with responsiveness to avoid ignoring valid but brief user input
• Monitor conversation logs to ensure important utterances aren’t being filtered out