Privatewhisper.ai: Don't type. Just talk. Private AI voice dictation
Private Whisper is a local and anonymous AI voice dictation tool for macOS, Windows, and Linux. Users can choose fully local processing, hardware-encrypted cloud, or anonymous cloud—audio is never stored or used for training. No account required, supports crypto payments. Features include hotkey dictation, multi-language support, noise adaptation, and use cases like coding, writing, and brainstorming.
Private Whisper — Local & Anonymous AI Voice Dictation
Don't type. Just talk.
Private AI voice dictation
Users choose — local, encrypted, or anonymous
Press a hotkey, speak, and clean text appears wherever you're typing — on macOS, Windows, and Linux. Run it fully on your device, inside a hardware-secured enclave, or in anonymous cloud; your audio is never stored or trained on. No account required.
Download Private WhisperAvailable on macOS, Windows, and Linux
From messy thoughts to clean text
Watch how Private Whisper transforms raw speech into polished text
Raw Audio
um so like I need to uh schedule a meeting with um Sarah about like the project timeline and stuff
Clean Output
I need to schedule a meeting with Sarah about the project timeline.
Choose your privacy level
Most dictation apps make the choice for you. Private Whisper lets you decide how your voice is processed — at every step.
Fully local
On-device transcription (Parakeet) and on-device cleanup (Local Gemma). Your audio and text never leave your device — works completely offline.
Nothing leaves your device
Confidential cloud
Cleanup runs inside a hardware-secured TEE enclave, encrypted on your device first — so neither PPQ nor any intermediary can read it.
Encrypted in a secure enclave
Fast cloud
Maximum speed and accuracy. Your audio is processed anonymously and is never retained or trained on by PPQ under any circumstance.
Never retained or trained on
No account, email, or KYC required
Pay anonymously with crypto
Only anonymous metadata — never your IP or identity
Everything you need to capture thoughts
No fluff, no complexity. Just the features that matter for getting ideas out of your head.
Get started without an account
Pay with crypto for no-KYC access and skip account setup for pay-per-use.
Lightning Fast
Hit your hotkey, speak, watch words appear instantly. No waiting, no friction.
One-hotkey capture
Instant start/stop without clicking around or breaking focus.
Handles accents and noise
Keeps up with different accents and less-than-perfect rooms.
Works in all your apps
One tool, endless possibilities. Private Whisper works seamlessly across all your favorite applications.
Built for every thinking moment
Whether you're coding, writing, or brainstorming - your voice is faster than your fingers.
LLM Prompting
Prompt AI models with your voice instead of typing long instructions
Code Comments
Dictate documentation and comments while you code
Brainstorming
Get ideas out of your head faster than you can type them
Content Creation
Draft blog posts, scripts, and articles by speaking naturally
Seamless onboarding
Download, configure, and start dictating in under a minute.
Simple, honest pricing
Free
Prefer to stay fully local? Run on-device transcription (Parakeet) and on-device cleanup (Local Gemma) and Private Whisper is completely free — no credits, no account. You only pay when you choose cloud processing.
Pay as you go and save
Average users spend only $2.50 per month!
No account required
Buy credits and load them into Private Whisper
Pay only for what you use
First 20 dictations are free
PPQ Accepts
Pay seamlessly with crypto or card
Deposit as little as $5 with card or 10 cents with crypto
Questions?
Everything you need to know about getting started with Private Whisper. 🤔
Private Whisper is incredibly cheap compared to competitors like Wispr Flow or Super Whisper. No subscriptions, pay per use. The average customer spends only $2.50 per month — and fully local, on-device transcription is completely free.
Yes. Choose Parakeet for on-device speech-to-text and Local Gemma for on-device cleanup, and your audio and transcripts never leave your device — it even works offline. This fully-local pipeline ships in the current public release. See our privacy page for details.
You choose your privacy level. Run everything fully on-device (Parakeet + Local Gemma) so nothing is sent anywhere; use the Tinfoil hardware-secured enclave so your cleanup is encrypted on your device and not even PPQ or Tinfoil can read it; or use fast cloud processing, where your audio is never retained or trained on by PPQ under any circumstance. Pay with crypto and you can use the service with no account, no email, and no KYC. See our privacy page to learn more.
Yes. Unlike most dictation apps, Private Whisper requires no account, email, or KYC, and offers a fully on-device mode where your audio never leaves your computer. You can also pay anonymously with crypto and pay only for what you use, with no subscription tying a recurring identity to your dictation.
No. PPQ never retains or trains on your audio under any circumstance. In fully-local mode nothing leaves your device at all. For cloud transcription, a default-on setting keeps your audio from being retained or trained on by our speech-to-text provider. We only record anonymous metadata (such as timestamp, audio duration, and language) — never IP addresses or personally identifiable information.
Private Whisper lets you choose at each step. For speech-to-text, Parakeet runs entirely on your device (free, offline, ~680MB one-time download), while Deepgram cloud offers maximum accuracy with audio that is never retained or trained on by PPQ. For cleanup, you can run Local Gemma on-device, use the Tinfoil hardware-secured enclave, or use fast cloud processing. Local modes keep everything on your machine; cloud modes trade that for a bit more speed and accuracy.
Dictation has been around for decades, but it always felt clunky — robotic output, constant misheard words, and no understanding of context. In the last year, AI models have leapt forward dramatically. Modern speech-to-text models are far more accurate across accents, background noise, and natural speech patterns. On top of that, large language models can now clean up raw transcriptions intelligently — removing filler words, fixing grammar, and formatting text the way you actually meant it. The result is dictation that finally feels natural, fast, and reliable enough to replace typing for real work.
Private Whisper is available on macOS, Windows, and Linux. It works in any app — just press your hotkey and start talking. Your transcribed text is pasted directly into whatever application is active, whether that's ChatGPT, Slack, Google Docs, VS Code, or anything else.
Private Whisper supports dozens of languages and accents. Local Parakeet transcription covers 25 languages, and cloud transcription supports most major world languages with auto-detect. English has the highest accuracy.
After installing Private Whisper, open the app settings to choose your preferred hotkey. The default is the globe key on macOS. You can change it to any key combination that works for your workflow.
When you press your hotkey, Private Whisper records your audio and transcribes it with a speech-to-text model — either locally on your device (Parakeet) or via the cloud (Deepgram). The raw transcription is optionally passed through a language model that cleans up filler words, fixes grammar, and formats the text — on-device (Local Gemma), inside a hardware-secured enclave (Tinfoil), or in the cloud. The polished result is pasted into your active app in a couple of seconds.
Ready to write 3x faster?
Write your next email, message, or prompt with Private Whisper.
Download Private WhisperAvailable on macOS, Windows, and Linux