Independent Mac software
Focused tools that keep your data on your Mac. The first is Local Dictate — it turns anything your Mac plays into text, with no audio sent to a server.
Apps
No suite, no subscription treadmill. Just the tools that earn their place in the menu bar.
Private, on-device dictation for meetings, videos and calls. Listen to what your Mac is playing, get a transcript, and save it as text. English and Romanian, with optional translation.
Local Dictate · macOS
Meetings, lectures, interviews, videos. Press Listen, and Local Dictate turns the system audio into text while you keep doing what you were doing.
First release in preparation. DMG, checkout and license portal links will open here when the store is ready.
What makes it different
Local Dictate is built around a simple constraint: the useful part should work without sending your audio to someone else's computer.
Capture what your Mac is already playing: calls, videos, webinars, voice notes. No extra audio driver.
OpenAI's Whisper large-v3 Turbo runs locally through CoreML, using the GPU and Neural Engine.
Automatic language handling for one language at a time, plus optional translation in both directions.
Teach it names, product terms and recurring mistakes. Replacements are case-aware and never chain unexpectedly.
Copy, clear, save as .txt, pause and resume. A−/A+ text sizing keeps long transcripts comfortable.
Sparkle handles downloads and relaunches. Check when you want from the Update button — no background nagging.
How it works
Local Dictate captures the audio your Mac is playing. Pause or stop whenever you like.
A local Whisper model turns each spoken segment into text as soon as there is a pause.
Copy the result, save a .txt file, or translate finished segments into English or Romanian.
Privacy by construction
No uploads, no transcription accounts, no analytics inside the app. The only network calls are for model downloads, optional update checks and license validation.
Read more in the privacy policy. Model downloads and license checks need a network connection; transcription does not.
Licensing
Buy Local Dictate once and activate it on up to five Macs at the same time. Your license key lives in the macOS Keychain, and you can deactivate old machines from the Polar customer portal.
Read the license terms or visit support if you need a hand.
FAQ
No. Audio is captured and transcribed on your Mac. Model downloads, optional update checks and license validation are the only network activity.
Yes. Every new install includes 10 minutes of active recording, with translation included. The trial works offline, needs no account, and stops counting when you pause. After it ends, existing transcripts can still be saved; a license key unlocks new recordings.
English and Romanian for transcription. Optional translation runs locally in both directions between the two languages.
Yes, once the speech model has been downloaded. Translation is optional and needs its own local models. License validation includes an offline grace period; you will only need a connection occasionally.
Up to five active Macs per license. Deactivating a Mac frees a slot. Use the Polar customer portal to see and remove old activations.
Not in the first release. Local Dictate captures system audio: what your Mac is playing. Microphone capture is not part of this version.
macOS 15 or later on Apple Silicon. The speech model takes about 1.5 GB; optional translation models add about 1.5 GB. System Audio Recording permission is required the first time you press Listen.
Polar emails the key after purchase and keeps it in your customer portal. You can paste the key into Local Dictate on each Mac.
Local Dictate includes Sparkle. Use the Update button or the app menu to check for a newer version; Sparkle downloads, installs and relaunches the app for you.
