### Answer You can use the pronunciation dictionary feature to guide how the agent pronounces specific words or symbols like "@". This lets you define custom phonemes for consistent pronunciation. [(1) ]() Here's an example of how to set it up when creating or updating an agent:

javascript
pronunciation_dictionary: [
  {
    "word": "actually",
    "alphabet": "ipa",
    "phoneme": "ˈæktʃuəli"
  }
]

You would add an entry for "@" with the phoneme for "at" (e.g.,

æt

in IPA) . Additionally, enabling normalize_for_speech can help with consistent speech synthesis for numbers, currency, dates, and similar text patterns. When set to

true

, it normalizes certain text to its spoken form before audio generation.

javascript
normalize_for_speech: true

Note: The pronunciation dictionary is currently only supported for English & 11labs voices.

​ To learn more, see the sources I considered: [[1] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

### Answer There are several agent settings that can affect voice detection and response latency: Responsiveness Controls how responsive the agent is. Value ranging from [0,1]. Lower value means less responsive agent (wait more, respond slower) , while higher value means faster exchanges (respond when it can) . If unset, default value 1 will apply. [(1) ]() STT Mode Determines whether speech-to-text should focus on latency or accuracy. Options are

fast

,

accurate

, or

custom

. Default is fast mode. [(2) ]() Interruption Sensitivity Controls how sensitive the agent is to user interruptions. Value ranging from [0,1]. Lower value means it will take longer/more words for the user to interrupt the agent. If unset, default value 1 will apply. [(1) ]() Denoising Mode Options include

noise-cancellation

or

noise-and-background-speech-cancellation

. Background noise can affect voice detection. [(2) ]() Things to try: - Increase

responsiveness

toward 1 for faster exchanges [(1) ]() - Set

stt_mode

to

"fast"

if latency is a priority [(2) ]() - Increase

interruption_sensitivity

toward 1 if the agent isn't picking up your speech quickly [(1) ]() - Try

noise-cancellation

denoising mode if you're in a noisy environment [(2) ]() If you're still experiencing issues, it could be related to network latency, microphone quality, or background noise on your end.

​ To learn more, see the sources I considered: [[1] Update Voice Agent](https://docs.retellai.com/api-references/update-agent) [[2] Create Voice Agent](https://docs.retellai.com/api-references/create-agent) ​

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.

Thanks for the feedback, glad I was helpful.

@AI Evy close this

### Answer I'm unable to close Discord threads—that's outside what I can help with as an AI assistant for Retell AI. If you have any other questions about Retell AI, feel free to tag me again!

message has been deleted

Feel free to tag @AI Evy with additional questions. I also appreciate feedback, it helps me improve.