Happy Voice lets you talk to your AI coding sessions instead of typing. It works through a voice gateway that processes your speech and intelligently routes it — either responding directly or forwarding instructions to your AI coding agent.

​
How it works

The voice pipeline has three stages:
  1. Speech-to-Text (STT) — Your voice is transcribed in real time
  2. Language Model (LLM) — An intermediate AI decides what to do with your message
  3. Text-to-Speech (TTS) — The response is spoken back to you
The voice agent acts as a smart bridge. It decides whether to:
  • Forward to your coding agent — For coding instructions like “refactor the auth module”
  • Reply directly — For conversational questions like “what’s the status?”
  • Call a tool — For app actions like “switch to session 2” or “approve the permission”

​
Supported languages

Voice works in multiple languages. The agent performs automatic speech correction for technical terms and programming jargon, which is especially useful for non-English languages where STT may misinterpret code-related terminology. Supported languages include: English, Chinese, Japanese, Korean, Spanish, French, German, and more.

​
Voice providers

Happy supports two voice providers:
ProviderDescription
Happy VoiceBuilt-in, works out of the box with the app — no setup needed
ElevenLabsThird-party voice AI, requires an ElevenLabs account
Happy Voice is the default and recommended provider. It’s built into the app and ready to use immediately. See Voice Setup for configuration details.