Give your keyboard
a nap.

Hold a hotkey and talk. Keynap transcribes on-device and types the text straight into whatever window has focus. No cloud, no upload — your voice never leaves the machine.

Coming soon to Microsoft Store See how it works
Typing

Let's move the kickoff to Thursday after

As far as you got
Speaking

Let's move the kickoff to Thursday afternoon and send the updated agenda to the team before the call.

3× faster Done — full message

Same amount of time. Speaking writes far more — and it lands in the focused app.

Why Keynap

Private by design, fast by default

Everything that matters about a dictation tool — without sending your voice anywhere.

01

Fully local. Nothing uploaded.

Speech recognition runs entirely on your device. No audio, no transcripts, no telemetry leave your computer — fit for confidential, medical and legal work.

02

GPU-accelerated transcription

Keynap auto-selects the fastest backend available: NVIDIA CUDA, then Vulkan, then CPU. On an RTX 4070 with CUDA it transcribes at around 33× real time.

03

Push-to-talk into any app

Hold your hotkey, speak, release. The text is injected into the focused window — editors, chat, browsers, terminals. It works wherever you can type.

04

Multilingual

Dictate in Japanese, English and many other languages. Pick your language in settings, or let it follow your system.

How it works

Three steps, no menus

Click where you want text, hold the hotkey, talk. Keynap does the rest.

STEP 01

Put your cursor anywhere

Click into any text field — a document, a chat box, a code editor, an address bar.

Notes
Ctrl Win Hold to record
STEP 02

Hold the hotkey and speak

Press and hold Ctrl + Win and talk naturally. A small floating HUD shows the live waveform and whether you are speaking or silent.

00:02.3 Speaking
STEP 03

Release — it types

Let go of the hotkey. Keynap transcribes on-device and types the text into the focused app, right where your cursor is.

00:04.1 Transcribing
Wrap up the notes and share them with the team.
Requirements

Runs on what you already have

A GPU makes transcription faster, but Keynap also runs on CPU. The speech model ships with the app — no separate download.

Minimum

Still works great

GPU None required runs on CPU
Memory 8 GB baseline
Pricing

Start free. Upgrade when you need more.

Core voice typing is free forever. Premium is a one-time purchase — a perpetual license that also includes every future premium feature.

Basic
Free

Free forever · no account

Get it on Microsoft Store Coming soon
Includes
  • Unlimited voice typing
  • Push-to-talk into any focused app
  • Fully local — works offline
  • Multilingual dictation
Premium Recommended
$29 one-time

Perpetual license · no subscription

Buy Keynap Premium Coming soon
Everything in Basic, plus
  • Transcription history, saved on your device
  • Copy and manage past transcriptions
  • All future premium features included
  • Buy once — yours forever
FAQ

Questions, answered

Does Keynap send my voice over the internet?

No. Recognition runs on your device and your audio stays on your computer. The speech model ships inside the app, so there is no model download — the only network use is getting and updating it through the Microsoft Store.

Do I need a GPU?

No. Keynap runs on CPU if no supported GPU is present. A CUDA or Vulkan GPU makes transcription dramatically faster, but it is optional.

Which languages are supported?

Japanese, English and many other languages. You can choose your language in settings or let it follow your system language.

Which apps can it type into?

Any app with a text field. Keynap injects the transcription into the focused window, so it works in editors, browsers, chat apps, terminals and more.

Is it available for macOS or Linux?

Not at the moment. Keynap is currently Windows-only (Windows 10 and 11, x64).

How much does it cost?

The core voice typing is free. Extra features such as transcription history come with Keynap Premium — a one-time $29 purchase with a perpetual license that also includes future premium features.