Skip to main content
WonkaChat can turn your voice into text, read responses back to you, generate images, and search the web for current information — when your organization and the agent you’re using support it.
Image generation and web search depend on both your organization’s settings and the specific agent you’re talking to. If you don’t see one of these for an agent that should have it, ask the agent’s builder or your administrator.

​
Speech to Text and Text to Speech

Configure voice behavior in Settings → Speech.

​
Speech to Text

1

Turn it on

In Settings → Speech, enable Speech to Text.
2

Use the microphone

In the message box, select Use microphone and speak — your words appear as text.
Additional options:
  • Auto transcribe audio — transcribe automatically without a manual stop
  • Auto send text — send your message automatically once transcribed (set to -1 to disable)
  • Decibel sensitivity — how sensitive silence detection is when auto-transcribing
  • Language and Engine — which speech-to-text language and engine to use

​
Text to Speech

1

Turn it on

In Settings → Speech, enable Text to Speech.
2

Pick a voice

Choose your preferred Voice from the dropdown.
Additional options:
  • Autoplay Latest Message — automatically read the assistant’s newest reply aloud
  • Audio Playback Rate — speed up or slow down playback
  • Use cloud-based voices — higher-quality voices served from the cloud instead of your browser’s built-in ones

​
Conversation Mode

Conversation Mode turns on a hands-free, back-and-forth voice exchange with the assistant — speak, listen to the reply, and keep talking without touching the keyboard. Enable it in Settings → Speech.

​
Image Generation

If the agent you’re talking to has image generation enabled, just ask for an image in your message (for example, “generate an image of
”) and it appears inline in the conversation.
Image generation is an agent capability, turned on per agent by its builder — it isn’t a setting you toggle yourself.
If the agent you’re talking to has web search enabled, it can look up current information instead of relying only on what it already knows. You’ll see a “Searching the web” indicator while it works, and results are read and summarized before the agent answers.
Web search is enabled per organization and per agent. Not every agent has it — check with the agent’s builder if you expect it and don’t see it being used.