Notes and drafts
Capture thoughts quickly in Apple Notes or any simple text field.
Native macOS dictation
Local multilingual speech-to-text for Mac.
Press Right Command, speak naturally, and your words appear in Notes, Pages, Word, Mail, and other everyday writing apps.
This is a Mac app. On a phone, use Share or AirDrop to send this page to your Mac, then download there.
Right Command
Local
11 primary + 2 experimental
Your language matters
Diktator! understands your language.
Lite has 11 primary languages, plus Arabic and Turkish as experimental languages that current tests recognize quite well. More languages will be added and existing language quality will improve over time.
Brand character
A little imperial. Still a calm Mac utility.
The Caesar illustration gives Diktator! a memorable voice, while the product experience stays small, quiet, and practical.
Mac first
Made for writing, not command lines.
Diktator! is meant for everyday text: notes, drafts, messages, letters, reports, and documents. Terminal commands and code editors are not the main use case.
Pages and Word
Dictate longer documents while keeping transcription local on your Mac.
Mail and messages
Use your voice for replies and everyday writing without opening a separate recorder.
How it works
One simple rhythm.
Press Right Command
Start recording with the Lite hotkey. Press it again to stop.
Speak naturally
The small capsule at the bottom of the screen shows a green wave while you speak.
Text appears
Diktator! transcribes locally and pastes the result into the app you were using.
More than speech recognition
From speech to readable text.
Diktator! does not simply record audio and hand it to a speech model. It adds a dictation layer designed for real writing: local processing, natural pauses, surrounding context, spoken formatting, and a cleaner final result.
Local on your Mac
Audio and text are processed locally. They are not sent to cloud transcription servers.
Adaptive pauses
The next advanced version is testing speech-rhythm handling so slower reading and faster speech can be split more naturally.
Context-aware transcription
Fragments should not feel like isolated scraps. The dictation layer is designed to keep nearby context for better continuity.
Live draft + final polish
A live draft and a final cleanup pass are in testing for the next advanced version, aiming for better punctuation, grammar, and flow.
Spoken formatting
Basic words such as comma, question mark, new paragraph, and capitalization can be treated as formatting rather than literal text.
Designed for real dictation
The goal is readable text that feels closer to what you meant to write, not a raw transcript that needs heavy repair.
How this differs from a simple speech model UI
A speech model turns audio into words. Diktator! manages the writing workflow around it: when to listen, how to split speech, how to preserve context, where to place the text, and how the result can be cleaned up after dictation. The public Lite build is the first local Mac release; the richer live draft and adaptive long-form layer are being tested for the next advanced version.
Speech to text
AI-style dictation, outside the chat box.
If you like voice input in ChatGPT, Perplexity, Grok, or other AI tools, Diktator! brings a similar speech-to-text rhythm to Mac writing apps: Notes, Pages, Word, Mail, and ordinary text fields. It is not text-to-speech; it turns your voice into text.
AI dictation for Mac
Press Right Command, speak, and let the transcript land where your cursor already is.
Compare the workflowSpeech to text in your language
Search phrases like voice to text, диктовка на русском, dettatura in italiano, and dictado en español all point to the same need.
See language phrasesРусская диктовка для Mac
Для Яндекса и русскоязычного поиска есть отдельная страница про голос в текст на русском языке.
Открыть русскую страницуAvailable now
Lite is useful from day one.
Lite is free in this release and is intended to remain a useful free version. It includes Right Command, one recommended local model, and automatic language detection.
Download version 0.2.8Coming later
Pro will add control, not pressure.
Pro is planned as a future one-time purchase for model and language choice, new local models, long-form dictation, optional user-consent audio capture workflows, local archive features, and automatic English translation.
See the Pro upgrade