🎙 Voice input for people who find typing hard · Mac & Windows

You speak. It types.

Speak, and text lands right at your cursor — no hands needed. Meetings take notes for you; select any text to rewrite or translate it. Recognition runs entirely offline on your machine: your voice stays yours alone.

Mac · App Store → Windows · Direct Download →

macOS 1.1.0 · Meeting notes / Select & rewrite / Live preview　·　Windows 1.3.6 (1.3.7 coming soon) · Meeting notes / Bring your own API key

Kaikou — Voice Input

Recognition runs offline· Audio never leaves your Mac

🗒 Meeting minutes write themselves· To-dos auto-extracted

Choose your platform

Two platforms, ready when you are

The same "speak and it types" experience, native to your device. Feature parity is maintained across both, each designed for its platform's conventions.

macOS · v1.1.0 · Freshly released

On the Mac, speaking is the shortcut

Download from the Mac App Store. Version 1.1.0 extends voice into everyday work: meetings, rewriting drafts, and watching recognition live — all just by talking.

Meeting notes with live recording preview
Select to rewrite / translate / paraphrase
AI minutes in 3 forms · Dialect correction · 30+ languages

Download from App Store Explore macOS →

macOS 14 or later · Basic dictation free forever

Windows · v1.3.6 · v1.3.7 coming soon

On Windows, the most complete meeting-notes loop

Direct download from this site. The app is free, no subscription required. Dual-track audio, speaker separation, minutes and to-dos all organized together.

Post-meeting Whisper refinement · Export minutes to Word / RTF
8 dialect-zone packs · Direct English typing
Free app · Bring your own API key (BYOK)

Download for Windows Explore Windows →

Windows 10 1809+ / 11 · Free app · AI uses your own key

📱

HarmonyOS Phone IME

Put "speak to type" into the HarmonyOS phone keyboard

In development · Coming soon

💻

HarmonyOS PC

The same experience on HarmonyOS tablets and PCs

In development · Coming soon

🤖

Android

Voice input for Android phones and tablets

In development · Coming soon

Meeting Notes

Just talk in meetings — the minutes write themselves

From audio capture, segmentation and speaker ID, to minutes, to-dos and export — one complete on-device pipeline. Your data never leaves the computer.

Dual-track audio: participants' sound is captured too (Windows)
Speaker separation with live paragraphed transcription
AI minutes: conclusions, decisions, to-dos and risks in one document
Task ledger: auto-extracted, tracked across meetings, with due reminders
Post-meeting Whisper refinement: re-transcribe the whole recording offline, live draft untouched

Post-meeting refinement: the offline Whisper model re-transcribes the entire meeting

Core Features

A complete toolchain from voice to polished text

Dictation, AI candidates, meeting notes and select-to-rewrite — use what you need.

🎙

Speak to type

Double-tap Command (Mac) or hold Right Alt (Windows), speak, and text lands right at your cursor.

📡

Offline speech recognition

An on-device recognition engine that works without the internet. Audio never leaves your machine and is discarded after processing.

👀

Live transcription preview New in 1.1.0

As you speak, the HUD shows recognized text in real time with auto-scrolling — confirm as you go, never wonder if you were heard.

✨

4-row AI candidates

Original, AI-polished, terminology, and Prompt — four candidate rows to pick from with one keystroke. Rewrite strength adjustable in three levels.

🗂

Meeting notes Both platforms

Just talk in meetings: live paragraphed transcription, speaker separation, AI minutes, auto-extracted to-dos — a complete record the moment you're done.

🖱