The problem
Typing slows you down. Every app switch breaks your focus. By the time you find the right tool, the idea is gone.
Dictate into any app at your cursor with one hotkey. Local Whisper and Parakeet models transcribe on your Mac or PC by default — no internet connection required, no audio upload. Faster cloud models exist too, but only on EU infrastructure and only if you switch them on.
7 days free · cancel anytime, effective at period end
Ownvox transcribes on-device — airplane mode included.
Transcription works offline
And most alternatives make things worse.
Typing slows you down. Every app switch breaks your focus. By the time you find the right tool, the idea is gone.
Cloud dictation apps send your voice to remote servers. Built-in OS tools are limited and clunky. Either way, you give up privacy or quality.
A tool that grows with you: on-device by default, cloud speed where you want it, and your own AI building blocks for the tasks you do every day.
Ownvox runs system-wide on macOS and Windows and turns speech into text in any application — your writing app, email, browser, an IDE — right where the cursor sits. No plugin, no copy-paste, no connection needed.
App and product names are trademarks of their respective owners and are used for reference only; they do not imply endorsement.
Typing and asking AI both break your flow. Ownvox closes both gaps, right at your cursor.
You already know the sentence, then spend seconds typing, fixing, and switching apps.
Hotkey, speak, done. Ownvox transcribes on your device and inserts text exactly where you're working, off the cloud by default.
An AI question means browser, chat window, copy, paste back, context gone.
Voice prompts from any app. The reply lands in your text. Your API key, your model, no window hopping.
Mode 2 up close: selected text is passed along as context automatically, and the answer lands right at your cursor, no copy-paste.
Browser, editor, notes app. One shortcut and Ownvox is listening.
Say what you need. Highlight text first and it is passed along as context automatically.
The model replies right into your text flow. No window hopping, no interruption.
Skills handle the rest, right at your cursor, no copy-paste.
How it works
Select text, press the hotkey, pick a skill from the radial menu. The AI output replaces your selection right inside the app you are working in. No copy-paste, no tool switching.
Hover any wedge to see the skill name. Click one to see the input it expects and the output it produces.
Skill menu
Hover or click a skill
Hover any skill to preview the input. Click to apply and watch it transform.
This is a static demo. In the real app the output replaces your selection in any text field.
Start from templates, shape them around your voice, and sort them with drag and drop.
All your skills in one place: cleanups, summaries, formatting, or your own prompt chains. Start from templates and shape them around your voice and tasks.
Arrange skills the way you actually work, frequent ones on top, rarely used below. A quick drag is all it takes for the radial menu to surface the skills you reach for in the flow.
Most dictation tools stream your audio to a vendor's servers and stop working without a connection. Ownvox transcribes on-device by default — on a plane, on locked-down networks, or with Wi-Fi off entirely. Cloud models exist, but only as an explicit opt-in on EU infrastructure.
Local Whisper and Parakeet models run on Apple Silicon and on Windows GPUs/NPUs. After a one-time model download, no connection is needed — dictate in airplane mode.
Audio is processed locally and is only ever uploaded if you deliberately enable a cloud feature. Confidential drafts, client notes and internal documents stay on your machine — the simplest possible GDPR story.
Hold your hotkey in any app, speak, and release. Punctuated, formatted text appears at the cursor — no window switching.
Whisper large models reach roughly 2% word error rate in English and about 4.3% in German — on your own hardware, with a custom vocabulary for your technical terms.
Want extra speed or AI post-processing? Optional cloud features run on EU infrastructure (Germany, France) with zero data retention — and a privacy switch turns them all off.
Dictate in over 50 languages with automatic language detection per dictation — offline included.
| Feature | Ownvox | Typical cloud tool |
|---|---|---|
| Where transcription runs | On your device | Vendor's servers |
| Works fully offline | Yes | No |
| Your audio leaves the device | Never, by default | Uploaded to the cloud |
| Works across every app | System-wide, at the cursor | Often a browser extension |
| GDPR posture | Local by default, EU cloud + DPA on opt-in | US cloud, varies |
| Platforms | macOS & Windows 10/11 | Varies |
Both plans unlock the same complete product. The only choice is whether you want to start flexibly or lock in the better yearly price.
incl. VAT
That's just €7.08/mo · incl. VAT
After the free trial, your subscription continues at the price shown above unless you cancel. You can cancel anytime, effective at the end of your current billing period.
Note: This is a subscription for digital content. You have a statutory 14-day right of withdrawal from the conclusion of the contract. During the 7-day free trial you can cancel anytime — you won't be charged as long as you cancel before the trial ends. More in the withdrawal policy.
One hotkey, one voice, one cursor. Text lands where you're working, instantly.
7 days free, then from €8.99/mo · cancel anytime, effective at period end