Open source · Pay once · macOS 26+

Speak it.
Author it.

Orathor turns your voice into text anywhere on your Mac, in any app and any text field. You think 3 to 5× faster than you type. Hold a key and talk: you watch the transcript appear live, and it lands at your cursor the moment you let go. The audio never leaves your Mac.

Download for macOS Star on GitHub

There's no account and nothing to set up. You'll be dictating within 30 seconds of installing.

Mail — New Message
Hold right speak release Double-tap to lock it on. That's the whole manual.
Built to disappear

Every feature earns
its place.

01 System-wide Works in every text field: Mail, Slack, Xcode, your terminal. If a cursor blinks there, Orathor types there.
02 Real-time You watch your words appear live in the overlay while you talk, and the finished text drops in at your cursor when you release the key. Built in Swift rather than Electron, so it launches instantly and barely touches the CPU.
03 Smart formatting Orathor handles punctuation, capitalization, and paragraph breaks for you, and it keeps up with technical terms and proper nouns.
04 Command mode Say "new line", "select all", or "delete that" to edit by voice without touching the keyboard.
05 History Everything you dictate goes into a searchable log that lives on your Mac and nowhere else.
06 Clipboard mode Dictate straight to the clipboard when you'd rather paste than insert. Handy for forms and picky apps.
Your numbers

It keeps count, quietly.

The main window shows how much you've dictated: words, sessions, speed, and time saved. Every number is computed on your Mac from your own history.

Choose your engine

Local by default.
Cloud when you say so.

Out of the box, Orathor uses Apple's on-device SpeechAnalyzer, so audio never leaves your Mac. If you'd rather run a cloud engine, bring your own API key.

Apple SpeechAnalyzer default on-device
Deepgram Nova your API key cloud
OpenAI Realtime Whisper your API key cloud
Pricing

Pay once. Or build it yourself.

The code is GPL-3.0 open source and free for everyone. The signed, ready-to-run app is one purchase, and it starts with a free week.

From source
Freeforever
Full on-device dictation, no trial, no key Cloud engines with your own API keys Build it in Xcode from the public repo
"Your voice stays on your Mac, not on someone else's server."

On-device processing is the default, not a paid upgrade. And since Orathor is open source, you can read exactly what the app does with your audio.

Questions

Asked, answered.

Why pay when the code is free?

You don't have to. Build it in Xcode and you get the full app with no trial and no key. The $29 buys the part that's genuinely work to ship: a signed, notarized binary that updates itself. It's also what keeps the project maintained.

Does my audio ever leave my Mac?

Not unless you tell it to. The default engine is Apple's on-device SpeechAnalyzer, and it works with wifi off. If you add a Deepgram or OpenAI key, audio for those sessions goes to that provider, and the app always shows which engine is active.

What happens when the trial ends?

Dictation pauses until you enter a license key. Your history and settings stay where they are, and nothing gets deleted. If you'd rather not pay, build from source and keep going.

Is it a subscription?

No. You pay $29 once and use it on three Macs, updates included.

Why does it need macOS 26?

Orathor is built on SpeechAnalyzer and Foundation Models, the on-device speech and language APIs Apple shipped in macOS 26. They're what make fast, private dictation possible without bundling gigabytes of models. Supporting older systems would mean shipping a different, worse app.

Which languages does it support?

Every language Apple ships an on-device model for. Orathor follows your system language by default and falls back to English when no local model exists. The cloud engines cover more languages if you need them.

Stop typing what you
could just say.

Download for macOS View the source

Free 7-day trial · $29 once, 3 Macs · macOS 26+, Apple Silicon