Voice

Speak naturally with Chalie and hear its responses out loud.

Speak and be heard

Chalie supports full voice interaction — speak your requests and hear the responses out loud. It feels much more like talking to a knowledgeable friend than typing at a text box.

Voice works through the web interface on desktop and mobile browsers.

Voice is built in

Voice ships with every install and upgrade — the on-device models come bundled, so there’s nothing to turn on and no internet connection needed to use them. There’s no Settings toggle to enable or disable voice; if it ever stops working, reinstalling Chalie is the fix.

The microphone button is ready to use in the chat input bar as soon as Chalie starts.

Speaking to Chalie (push-to-talk)

Press and hold the microphone button to record your message, then release to send it. Chalie transcribes your speech on-device and responds as normal.

Try it with anything you’d normally type:

  • “What’s on my calendar tomorrow?”
  • “Set a timer for 20 minutes.”
  • “Read me my latest emails.”

Audio up to 10 minutes long is supported in a single recording.

Hearing Chalie’s responses

Chalie reads its responses out loud using text-to-speech. For short answers it responds directly; for longer content like email summaries it reads a concise version aloud.

Voice and speed are not currently adjustable through the interface.

Voice models

Both speech recognition and synthesis run entirely on your device — your voice data never leaves your machine.

  • Speech-to-text: Moonshine (Whisper-family ONNX), running locally
  • Text-to-speech: Kokoro v1.0 ONNX, running locally

The models ship with every install and upgrade — there’s nothing extra to download.

Accessibility

Voice mode is a good option for anyone who prefers not to type or finds keyboard input difficult. Chalie handles natural speech patterns, filler words, and mid-sentence corrections — you can say “actually, I meant…” and it will adjust.