100% Offline · No cloud · No data sent

Your voice.
Your words.
Your machine.

VoiceScribe transcribes speech to text in real time — entirely on your Windows PC. No internet required. No subscriptions to cloud APIs. Your audio never leaves your device.

Download for Windows View pricing →

Real-time transcription · Powered by Zipformer / Sherpa-ONNX

What you get

Everything you need,
nothing you don't.

Built for writers, developers, journalists, and anyone who thinks faster than they type.

Fully Offline

All ASR processing happens locally via Sherpa-ONNX. Your voice data never leaves your PC.

Real-time VAD

Silero Voice Activity Detection starts and stops transcription automatically — no manual push-to-talk needed.

AI Post-Processing

Optional LLM cleanup: punctuation, formatting, translation. Works with local Ollama or OpenAI.

System Tray App

Lives quietly in your taskbar. Assign a global hotkey and start dictating anywhere — in any app.

Multilingual

English and Russian UI. Multiple ASR models available. Translate output to any language via LLM.

No Admin Required

Installs to AppData — no admin privileges needed. Clean uninstall via standard Windows Add/Remove.

File Transcription

Drop an audio or video file (MP3, WAV, MP4, MKV…) into the History page — VoiceScribe transcribes it offline using the same local ASR pipeline. No real-time recording required.

How it works

From voice to text
in seconds.

01

Download & Install

Run the installer — no admin rights needed. VoiceScribe installs to your user AppData folder and adds a system tray icon.

02

Choose your microphone and model

Select your input device. Pick an ASR model (Zipformer English, Russian, or multilingual). Download runs once in the background.

03

Press hotkey → speak

Assign a global hotkey. Press it from any window — VoiceScribe listens, detects speech with VAD, and transcribes in real time.

04

Text is inserted instantly

Transcript is typed directly into your active window — no copy-paste. Optionally pass through AI for cleanup or translation first.

05

Or transcribe an existing file

Have a recorded meeting, podcast, or video? Upload it in the History page — VoiceScribe processes the audio offline and saves the transcript as a regular entry, ready to search or re-process.

Not happy with the result? Switch the model and re-transcribe.

Every recording and imported file keeps its audio saved locally. Go to Settings → Models, pick a different model — say, upgrade from a small Zipformer to Whisper Large — then open any history entry and hit Re-transcribe. No need to record or import again.

Pricing

Simple, honest pricing.

Start free. Upgrade when you need more.

Pro · Monthly

$3 / month

Cancel anytime. No hidden fees.

  • Unlimited transcription
  • All ASR models
  • AI post-processing (Ollama / OpenAI)
  • Multilingual output & translation
  • Up to 2 devices per license
  • All future updates

Free tier: 30 min/day · no card required

Opening secure checkout…