FREE v0.1.0

Voice to text,
anywhere on Windows.

Press a hotkey, speak, and your words appear in any app. Powered by local AI — no internet, no fees, no data collection.

Any Application
Local AIGPU AcceleratedPrivacy FirstFree Forever

Three steps. That's it.

Press your hotkey

Ctrl+Shift+Space (customizable)

Speak naturally

VAD detects your voice automatically

Text appears instantly

Pasted into the focused app

Everything you need to dictate

Local speech recognition, GPU acceleration, text-to-speech, and more — all in one lightweight app.

Powered by Whisper

Six AI models from tiny (75 MB) to large (3.1 GB). Choose your speed-accuracy tradeoff. Every word processed locally through whisper.cpp — no data ever leaves your machine.

Your data stays local

No accounts. No telemetry. No cloud dependency. Voice and transcriptions are processed entirely on your hardware. Delete the app and everything goes with it.

GPU Accelerated

CUDA support for NVIDIA GPUs. 5–10x faster transcription with real-time results.

Read Aloud

Six TTS voices with adjustable speed. Streaming playback through OpenAI.

Smart Cleanup

LLM post-processing with custom prompts and a domain-specific dictionary.

Cloud Fallback

Optional OpenAI, Mistral, or OpenRouter integration when you want it.

Choose your edition

Both editions are identical in features. The difference is how speech recognition runs.

CPU Edition

~22 MB

Works on every Windows PC. Whisper runs on your processor — no special hardware needed.

Best for: any PC without an NVIDIA GPU, or older/low-end cards.

Download CPU Edition

CUDA Edition

~427 MB

Optimized for modern NVIDIA GPUs (GTX 900+). 5–10x faster transcription. Bundles the CUDA runtime.

Best for: PCs with an NVIDIA GPU and up-to-date drivers.

Download CUDA Edition

Not sure? Start with CPU — you can switch to CUDA later without losing settings.

License: Proprietary. Source code is not available for redistribution, modification, or derivative works.

Pricing: Currently free. Future updates may introduce paid licensing.

Disclaimer: Provided "as is" without warranty. Authors are not liable for any claims or damages.

Cloud API usage is subject to each provider's terms. Local models use OpenAI Whisper weights via whisper.cpp.

Up and running in minutes

1

Install

Run the .msi installer. Creates a Start Menu shortcut. No admin rights required for per-user install.

2

Download a model

Open Settings > Models. Pick a model (we recommend Distil Large v3) and click Download. Runs entirely offline.

3

Set up your mic

Go to Settings > Audio. Select your input device, hit Test to verify, and adjust VAD sensitivity if needed.

4

Start dictating

Press Ctrl+Shift+Space. Two modes: push-to-talk (hold hotkey) or toggle (press once to start, again to stop).