MADE FOR LINUX DESKTOPS

Have a thought?
Just say it.

Press a shortcut and speak. Doubao Say turns your voice into text in the app you are using.

◉ Omarchy / Hyprland Native X11 Open source

Keep up with your thoughts.Make typing feel lighter.✳Put your hands to better use.
WHY DOUBAO SAY

From voice to text.
One smooth flow.

Keep your familiar desktop. Doubao Say puts voice input within reach whenever you need it.

01 / TRIGGER
⌨

Press and speak

Tap to start and tap again to finish, or hold to talk and release when done. Choose a preset shortcut or record your own.

02 / TRANSCRIBE
〰

See words appear

Live text appears in a floating overlay. Send the result to the active app, or try it inside Doubao Say first.

03 / REFINE
✳

Optional text polish

Connect your own OpenAI-compatible model to clean up filler words and punctuation. If polishing fails or times out, the original text is used.

04 / CHOOSE
◎

Three ways to transcribe

Use your Doubao account by default, bring a key for Volcengine Seed ASR 2.0, or dictate in US English with Deepgram Nova-3.

A SIMPLE FLOW

Let your voice
do the typing.

Stay in your app. Whether you are writing, chatting, or taking notes, speak wherever the cursor rests.

Read the full installation guide
  1. 01

    Choose a service and microphone

    Sign in with a Doubao web account by default, or configure Volcengine Seed ASR 2.0 or Deepgram Nova-3 for English.

  2. 02

    Press your shortcut

    Start and stop recording from the floating overlay. Adjust the microphone and shortcut in Settings.

  3. 03

    Text reaches your app

    Hyprland supports clipboard paste or optional direct input. Native X11 supports clipboard paste.

BUILT FOR YOUR DESKTOP

Made for Linux.
Made to flow.

Use it as a standalone GTK4 app or an Omarchy plugin. Settings, microphone selection, and first-run setup live in the app.

OmarchyHyprlandWaylandX11
Doubao Say running on an Omarchy Linux desktopReal app interface · Omarchy Linux
READY WHEN YOU ARE

Your next thought?
Say it out loud.

Get a release for your CPU architecture and Python version from GitHub, or install from source using the guide.