fn VOICE INPUT
Dictate at the cursor.
Stay in the text field you are using. Choose press and hold, or tap once to start and again to stop.
A simple thought, written down.
Let's talk about the new design on Friday.
SPEAK YOUR MIND.
Speak. It lands where you type.
Press Fn to dictate at the cursor. Double press Fn and say what you need. Ask a question, draft text, or follow up without selecting anything; select text to rewrite or translate it. Built as a native Mac app after I tried alternatives when Typeless reduced its allowance.
macOS 14+ · Native Mac app · Free app · 5.5 MB installer
fn × 2 VOICE AGENT
Say what happens next.
Double press Fn to ask a question, draft text, or follow up—no selection needed. Select text when you want it rewritten or translated.
VOICE AGENT EXAMPLE
“Make that sentence shorter.”
Small Memory, clear control
Add names, projects, and terms by hand, review suggestions from pasted text, or confirm a correction after dictation. Edit or delete saved items whenever you need.
See how you use your voice
Usage shows daily, weekly, and monthly dictation totals, speaking time, top apps, and Voice Agent tasks. Daily totals stay on this Mac without text or audio. Pause or clear them in Settings → Privacy.
Free app, your own model account
SayKuku is free. Its source is available on GitHub under Apache 2.0. It uses Qwen3.8 Omni Flash models to work directly with audio, rather than a fixed ASR-then-LLM pipeline. The current build uses Qwen cloud processing for voice input and Voice Agent. Audio and the context you allow are sent to your selected region. History, Memory, and optional recordings are saved in the app’s Application Support folder on your Mac; your API key is stored in macOS Keychain. Model use may incur charges.
Download the signed and notarized Mac app on the download page. The interactive demo on the Chinese home page uses preset text and does not access the microphone or call a model.