WHY SONAVI
Sonavi turns speech into text directly on your Windows PC. It combines local transcription, clear privacy controls, and a fast system-wide dictation workflow. This lets you write naturally without making your voice someone else's data.
Private by default. Built in Austria, Europe. Designed for Windows.
TEN REASONS
Local-first privacy meets practical, everyday Windows dictation.
On-device speech recognition
Sonavi processes your speech directly on your Windows PC using locally installed transcription models. Your voice does not need to be sent to a cloud transcription service to become text.
Local processing creates a clear privacy boundary: the words you dictate are handled on the computer in front of you. That makes Sonavi a strong fit for confidential notes, internal documents, early ideas, and everyday professional communication.
Audio discarded after processing
Microphone audio is processed for transcription and discarded after inference. Sonavi does not save a library of your voice recordings as part of normal dictation.
Your microphone should help you write, not create another archive of sensitive material. Sonavi's normal dictation flow keeps audio transient and does not write it to transcription history.
No transcript upload by default
Transcript content stays on your device by default. Sonavi does not upload the text you dictate simply to provide transcription.
The result is text you can use in the application you chose. Sonavi does not require your transcript to become cloud content, an advertising profile, or training material. Any optional content-sharing experience must be clearly explained and require explicit consent.
Private Mode and explicit opt-ins
Private Mode stops history recording and telemetry queuing immediately. Separate settings let you decide whether anonymous product analytics or optional improvement data may be shared.
Privacy should not depend on hidden defaults or a difficult account process. Sonavi places the relevant controls inside the application and keeps optional analytics separate from transcript and audio sharing choices.
Windows user-bound encryption
When history is enabled, transcripts are stored locally and encrypted for the current Windows user. You can delete individual entries, clear everything, or use Private Mode to avoid creating history at all.
Sonavi uses Windows Data Protection API protection for saved transcript history. Retention and deletion controls keep you in charge of how long that local record remains useful.
Austrian product development
Sonavi is developed in Austria, a country in the heart of Europe, with a product philosophy shaped by user control, data minimization, and transparent choices. Privacy is part of the architecture, not a marketing add-on.
Our Austrian development roots are reflected in how the product behaves. Sonavi is designed to minimize the data required for dictation and to keep sensitive voice and text processing close to the user.
System-wide hotkey dictation
Use a global hotkey, speak naturally, and insert the finished text into the active application. Sonavi fits into your existing workflow instead of asking you to write inside another cloud editor.
Emails, documents, messages, notes, and business tools remain where they are. Sonavi runs from the Windows tray and makes dictation available across compatible text fields, with clear recording, processing, completion, and error states.
Core transcription without the cloud
After the application and a transcription model are installed, core dictation can run without an internet connection. That is useful while travelling, on restricted networks, or whenever connectivity is unreliable.
Because speech recognition runs locally, an internet outage does not have to interrupt the core speech-to-text workflow. Internet access is still needed for model downloads, updates, feedback, licensing checks where applicable, and optional online services.
Local English and multilingual models
Sonavi supports a selection of local transcription models, including multilingual options. Choose a lighter model for responsiveness or a more capable model when accuracy matters most.
Different computers and workflows need different trade-offs. Sonavi's local model approach gives you a practical choice instead of forcing every dictation through one fixed remote service.
Fast workflow, transparent status
Sonavi turns natural speech into usable text with a hotkey-first workflow, clear status feedback, and automatic insertion. Local time-saved insights help show the value without exposing what you said.
The experience is designed to stay out of the way: start, speak, review the result, and continue working. Performance metrics use counts and timings rather than transcript text, with validated hardware acceleration and a reliable CPU fallback where supported.
WHAT STAYS LOCAL?
Sonavi makes the default behavior visible, so you can understand what stays on your PC and what needs a connection.
Sonavi's core speech-to-text processing runs locally after the required model is installed. The application may use an internet connection for model downloads, software updates, licensing or entitlement checks where applicable, user-submitted feedback, and optional analytics or improvement programs enabled by you.
READY WHEN YOU ARE
Sonavi gives you the speed of modern speech recognition without requiring your everyday thoughts, drafts, and conversations to become cloud data. Install it on Windows and make your voice part of the workflow you already trust.
Windows 10 and Windows 11 · Local transcription · Multilingual model options
