Coming February 2026

Voice to Text.
100% Offline.

AI-powered transcription that runs entirely on your computer. Your recordings never leave your machine. No cloud. No subscriptions. Just privacy. From BuddyIT — Kraków, Poland.

Learn More
14-day free trial
No credit card
Windows 10/11
100%
Offline Processing
0
Data Sent to Cloud
High
Accuracy (Whisper AI)
∞
Transcriptions
// FEATURES

Everything you need.
Nothing in the cloud.

Professional transcription powered by Whisper AI, running entirely on your hardware.

True Privacy

Your audio never leaves your computer. No uploads, no cloud processing, no third-party access. Ever.

NVIDIA & AMD Support

Dual-backend architecture: CUDA for NVIDIA GPUs, DirectML for AMD & Intel. Auto-detected, manually overridable.

Whisper AI Powered

State-of-the-art speech recognition. Multiple model sizes available — from fast to ultra-accurate.

No Subscription

Pay once, use forever. No monthly fees, no usage limits, no surprise charges. Your license, your software.

Multi-language

Supports English, German, Spanish, French, Polish and more. Automatic language detection included.

EV Code Signed

Digitally signed with Extended Validation certificate. Windows trusts BuddyVox — no SmartScreen warnings.

EV Code Signed

BuddyVox is signed with an Extended Validation certificate — the highest level of code signing trust. Windows recognizes BuddyVox as verified software from a verified publisher. No SmartScreen warnings, no "Unknown publisher" alerts.

Verified publisher identity
No SmartScreen warnings
Tamper-proof installer
// HOW IT WORKS

Three steps to transcription

1

Record or Import

Hit record to capture audio directly, or import existing audio/video files.

2

AI Transcribes

Whisper AI processes your audio locally on your GPU. Fast and accurate.

3

Export Anywhere

Copy to clipboard, save as TXT, SRT subtitles, or export to your favorite apps.

// PREVIEW

See BuddyVox in action

A sneak peek at the interface. Clean, focused, efficient.

BuddyVox v1.0

Screenshot coming soon

// PRICING

Simple, honest
pricing.

No subscriptions. No hidden fees. Pay once, use forever.

Free Trial

$0 14 days

Try before you buy. Full features, no credit card required.

  • All transcription features
  • All Whisper models
  • GPU acceleration
  • Export to TXT, SRT
  • 14-day limit

Professional

$49.99 one-time

For power users and small teams. Priority support included.

  • Everything in Personal
  • Priority email support
  • Batch processing
  • Custom export templates
  • Free updates (v1.x + v2.x)
  • 2 license resets per year
// ABOUT

Built in Kraków,
for the world.

BuddyVox is developed by BuddyIT, a small software studio in Kraków, Poland. We build tools that respect your privacy and work offline by default.

We started BuddyVox because we needed a transcription tool that didn't require uploading sensitive recordings to the cloud. Existing solutions either required internet access, had subscription fees, or raised privacy concerns.

So we built what we wanted: fast, accurate, offline transcription. One-time purchase. No data collection. No BS.

— Adrian & Edyta, BuddyIT

// FAQ

Frequently Asked Questions

Do I need an internet connection to use BuddyVox?

Only for the initial activation. After you activate your license once, BuddyVox works completely offline. No internet required for transcription, ever.

We designed it this way because we understand that privacy-conscious users don’t want their transcription software “phoning home.” Once activated, your copy of BuddyVox is truly yours.

What GPUs are supported?

BuddyVox supports both NVIDIA and AMD GPUs:

  • NVIDIA: GTX 1060 or newer (CUDA backend)
  • AMD: RX 5000 series or newer (DirectML backend)
  • Intel: Arc GPUs (DirectML backend)

For best performance, we recommend at least 6GB of VRAM. The larger Whisper models (medium, large) require more VRAM for optimal performance.

CPU-only mode is also available, but significantly slower than GPU acceleration.

How accurate is the transcription?

Very accurate — BuddyVox uses OpenAI’s Whisper model, which achieves near-human accuracy on many benchmarks.

Accuracy depends on several factors:

  • Audio quality: Clear audio = better results
  • Model size: Larger models are more accurate
  • Language: English has the highest accuracy, but 90+ languages are supported

For professional use, we recommend the “medium” or “large” model. For quick drafts or notes, the “small” model offers a good balance of speed and accuracy.

Can I use BuddyVox for commercial purposes?

Yes, absolutely. Your BuddyVox license permits commercial use. Transcribe client meetings, create subtitles for videos, process interviews — it’s all allowed.

The only restriction is that you cannot redistribute the software itself. One license = one user on one machine.

For teams or organizations needing multiple licenses, contact us for volume pricing.

What if I get a new computer?

Your license is bound to your hardware. If you get a new computer, contact our support for a license reset.

This is handled manually to prevent abuse, but we’re happy to help legitimate users. Just email us at [email protected] with your license key and we’ll reset it within 24-48 hours.

We understand hardware changes happen — we won’t make you buy a new license just because you upgraded your PC.

What's your refund policy?

30-day money-back guarantee, no questions asked.

If BuddyVox isn’t for you, just email us within 30 days for a full refund. We’ll deactivate your license and return your payment.

We offer this guarantee because we’re confident in our product. We’d rather have happy customers who choose to stay than unhappy customers who feel trapped.

Contact [email protected] for refund requests.

Ready to transcribe
with privacy?

Start transcribing with complete privacy. No cloud uploads, no subscriptions, no compromises.

Contact Sales