Can I dictate into any app?

Yes. Speech Dock uses a global hotkey and inserts the result into the active field where system input is supported.

Do I need separate API tokens?

No. Normal dictation does not require connecting to or paying for an external recognition service separately. The site uses the network only for installation, updates, checkout, license and support.

Which Linux versions are supported?

The main install instructions cover Ubuntu 22.04+, Fedora 38+, Debian, Arch and openSUSE. Wayland auto-paste uses ydotool, X11 uses xdotool.

What permissions are needed on macOS?

macOS asks for Microphone and Accessibility access so the app can listen for the hotkey and paste text into the active window.

No, a GPU is not required. On supported hardware acceleration may reduce processing delay, but exact numbers depend on your device.

How much disk space does the app need?

The installer itself is small. On first run the app picks components to match your computer — more accurate recognition on powerful machines, a lighter setup on weaker ones. It usually takes about 3 GB, less on low-end hardware — from roughly 1.5 GB. Everything stays on your device and works offline.

Where does my voice go?

Dictation is processed locally. We do not promise that the app never uses network: it is needed for the initial component setup, updates, checkout, license and support requests.

Are audio recordings saved?

Normal dictation does not require sending audio to a server. Local history and custom dictionaries stay on your device and are controlled in the app.

What data does the website process?

The website processes data required for purchase, license, receipts, support and consent management. Details are in the privacy policy.

Which cards do you accept?

The Russia MVP accepts Russian-issued bank cards. Payments are processed via YooKassa, and a 54-FZ receipt is sent by email.

Where do I get the license after checkout?

After successful payment the website issues a license and opens a license cabinet through a protected email link.

How many devices does one license cover?

Depends on the plan: Free — 1 device, Pro — up to 3, Lifetime — up to 5 simultaneously active. Manage and deactivate devices in your dashboard.

Refunds are governed by the terms of use. Check the current legal documents before purchase.

Is there a Windows version?

Yes — there's a dedicated Windows install guide (.exe installer or portable .zip). The current version status is shown right on the download page.

What is the default hotkey on Windows?

The default is Ctrl+Alt+Space. You can change it under Settings → Hotkeys if it conflicts with another app.

Will file transcription be available?

Capabilities depend on the tier and current app version. The homepage does not promise future features before implementation is ready.

Offline speech-to-text for Linux, macOS and Windows

A thought that becomes text instantly.

Download free Compare options

Linux · macOS · Windows · ~3 GB on first run

Open-source core·54-FZ receipts (RU)·Local history

Speech Dock

→

00:04

5× faster than typing

Say a sentence in seconds.
Typing it takes a whole minute.

The same text. On the left — by voice, on the right — by hand. Press Space to replay.

Voice

0wpm

dictating…

Keyboard

0wpm

typing…

Per independent Stanford research (2016): speech is on average 3× faster than typing. Source

6.4 h

saved on typing per week

at 2 hours of text per day

220

words per minute by voice

vs 40 on a keyboard

97%

recognition accuracy

in Russian and English

0 bytes

leaves for the cloud

processing stays local

Voice input that works as part of your system

Speech recognition, text clean-up, a personal dictionary and tuning to your computer — it all runs right on your device, with nothing sent to the cloud.

Your voice stays on your device

Recognition and text processing happen right on your device, with nothing sent to the cloud. Medical, legal and any other confidentiality stays with you.

All your recordings stay protected

History, text and settings are stored on your device in encrypted form, and the access key stays only with you.

Recognition that understands your language

Understands natural speech across many languages — colloquial phrasing, professional terms and sentences that mix words from different languages.

Turns speech into ready-to-use text

Helps shape what you said: removes filler, adds punctuation and paragraphs, and delivers the result as dictation, a note, a message or a to-do list.

Knows your names, brands and terms

A personal dictionary accounts for the names of colleagues and clients, product names and professional terms during recognition, not just after.

Adapts to your computer on its own

The app finds the right balance of quality and speed for your device and keeps the result from getting worse.

Runs locally on Linux, macOS and Windows.

Developers

Works everywhere
you write.

Scroll down — cards move sideways

debounce.js — ~/projects/utilsREC

00:04

01// User holds the hotkey and dictates
02function debounce(fn, delay) {
03  let timeoutId;
04  return (...args) => {
05    clearTimeout(timeoutId);
06    timeoutId = setTimeout(() => fn(...args), delay);
07  };
08}
09
10// Returns a function that delays the repeated call

Developers

Prompt by voice

Understands syntax, libraries and frameworks — dictate code and prompts straight into the editor without fixes.

Works in

VS CodePhpStormCursorZed

Team · #symphony

katya 10:02

robert 10:19

Hi team! This is the final week of the project — let's keep everything under control. @katya please share the status.

Message #symphony · hold to dictate

Customer support

Reply by voice

Dictate a reply and it becomes a tidy message to the customer — no switching between windows.

Works in

ZendeskSlackTelegram

Agreement — Docs

Service agreement

The parties have agreed on the following:

The contractor delivers the services by the 30th, and payment is made within five business days of signing the acceptance act.

Dictating…

Lawyers, doctors, teachers

Documents by dictation

Contracts, medical notes, lecture notes — speak your thought and it lands in the document as ready paragraphs, never leaving your device.

Works in

WordGoogle DocsNotion

Mail — Inbox

To Michael Robbins

Subject Marketing proposal

Hi Michael!

Please take a look at the proposal — I highlighted three key points: targeting, the Q3 budget and launch timing.

Dictating…

Business email

Emails by voice

Say the gist — the app shapes it into a ready-to-send email in the right tone.

Works in

GmailOutlook

‹AnyaA

today

Are you going to the meeting today?

We start at 9 😅

yes! running a bit late, will be around 9:30

Message

Personal messages

Tone to match the person

Professional for work, casual for close ones — the message style adapts on its own.

Works in

WhatsAppTelegramDiscord

01 / 05

What the app looks like.

One dock. No menus, no windows, no tabs. Everything one keystroke away.

Speech Dock — Main screen

Ready to dictate. A minimal dock with settings, history, and copying.

Privacy

Privacy without promising “no network ever”.

Voice input runs locally, but the website and purchase flow still use normal network services. These zones should stay separate.

Learn about privacy

Stays on your computer

Dictation audio and intermediate STT processing
Local history if you enable it in the app
Custom dictionaries and work terminology

May use the network

Initial download of recognition components and updates
Checkout, receipt and license via the website and YooKassa
Optional support requests or legal forms

Speech Dock vs
the usual options.

How it compares with what you'd actually consider for voice input — cloud services and built-in dictation.

	Cloud services ?	Built-in dictation ?
Your data never leaves your computer	—	depends on OS
Works in any app	via extensions	limited
Understands many languages and your terms	depends	basic
Shapes ready-to-use text: punctuation, paragraphs	depends	—
Works offline	—	yes, but basic
One-time purchase, no subscription	subscription	free, but basic

Choose your plan.

Pro Yearly — save 25%. Lifetime — one payment, forever.

Free

forever

30 minutes of speech per month, up to 5 minutes per recording
Auto-punctuation and capitals
Text cleanup for your task
Export to TXT
1 device

Download

Pro Monthly

every month

Unlimited dictation
Transcribe audio and video files
Custom terms and names dictionary
Export to PDF and Markdown
Local history encryption
Up to 3 devices

−20%

Pro Yearly

$48

per year · $4/mo

Save $12 per year
Everything in Pro Monthly
Priority support

Lifetime

$99

once, forever

Everything in Pro Yearly
Bring your own models
Up to 5 devices
All future updates

Russian-issued cards only · NPD (self-employed) receipt from YooKassa via email · 14-day refund window

International billing (USD and EUR) coming soon. For now, payment is accepted in RUB via YooKassa.

System requirements and tradeoffs.

Local processing depends on memory, disk and first-run setup. A GPU is optional, but supported accelerators may reduce processing delay.

Linux

OSUbuntu 22.04+, Fedora 38+, Debian, Arch, openSUSE — best-effort
Architecturex86_64
RAM8 GB minimum · 16 GB recommended
DiskThe required components download during first setup; keep several GB of free space
GPUNot required; supported accelerators may reduce processing delay
NetworkNeeded to download the required components, plus updates, checkout and license
Integrationydotool (Wayland), xdotool (X11)

macOS

OSmacOS 11.0 Big Sur or newer
ArchitectureApple Silicon · Intel x86_64
RAM8 GB minimum · 16 GB recommended
DiskThe required components download during first setup; keep several GB of free space
GPUNot required; Metal may be used on supported Macs
NetworkNeeded to download the required components, plus updates, checkout and license
IntegrationAppleScript, Accessibility API

Windows

OSWindows 10 or Windows 11
Architecturex86_64
RAM8 GB minimum · 16 GB recommended
Disk~300 MB for the app; the required components download during first setup, keep several GB of free space
GPUNot required; NVIDIA — CUDA, AMD/Intel — DirectML
NetworkNeeded to download the required components, plus updates, checkout and license
IntegrationWin32 API (global hotkeys, auto-paste, clipboard)

Ready to try?

Download the app, complete first setup and fetch the required components for local recognition.

Download for Linux Download for macOS Download for Windows

Mobile app (Android/iOS)

One email for Android and iOS. Unsubscribe in one click.

Frequently asked.

Speech recognition runs on your computer. After the initial component download, voice input does not require a cloud recognition service for every dictation.

Get notified of major releases.

I consent to data processing per Privacy Policy . Emails — only for major releases, unsubscribe in one click.

A thought that becomes text instantly.

Say a sentence in seconds.Typing it takes a whole minute.

Voice input that works as part of your system

Your voice stays on your device

All your recordings stay protected

Recognition that understands your language

Turns speech into ready-to-use text

Knows your names, brands and terms

Adapts to your computer on its own

Works everywhereyou write.

Prompt by voice

Reply by voice

Service agreement

Documents by dictation

Emails by voice

Tone to match the person

What the app looks like.

Privacy without promising “no network ever”.

Stays on your computer

May use the network

Speech Dock vsthe usual options.

Choose your plan.

System requirements and tradeoffs.

Linux

macOS

Windows

Ready to try?

Frequently asked.

Get notified of major releases.

Say a sentence in seconds.
Typing it takes a whole minute.

Works everywhere
you write.

Speech Dock vs
the usual options.