VoxBar™
100% Private AI Voice Transcription
Scroll for Demo ↓
100% Private AI Voice Transcription
No APIs. Zero Cloud. 100% Local.
Introducing VoxBar. A customizable UI for private, offline transcription—built for privacy, speed, accuracy, and control.
From vibe-coding and drafting emails, to capturing lectures and Teams or Zoom meetings via mic or system audio, VoxBar Pro transcribes in real time so your words become usable text instantly.
Explore our top-tier offline engines today, and experience what it's like to talk your ideas into reality.
The same S-tier Voxtral engine, now accelerated by Apple Metal GPU. No Docker. No cloud. No subscriptions. Just double-click and talk.
Native Apple Silicon performance. M1, M2, M3, M4 — all supported.
On Mac, VoxBar Pro runs without Docker. Pure native C engine.
Double-click the launcher. It builds the engine, downloads the model, and launches — all automatically.
Requires Apple Silicon Mac (M1+) with 8GB+ RAM. 16GB recommended.
Optimized for Windows. Every tier runs 100% locally. No cloud. No subscriptions. One-time purchase, lifetime license.
Performance
GLM-ASR Nano 1.5B • 17 Languages • 4GB VRAM
one-time
🔥 LAUNCH: $19.50 with code EARLYBIRD
VRAM: ~4GB • NVIDIA only • 17 languages
Get GLM — $39Performance
Nemotron 0.6B • English • 4.8GB VRAM
one-time
🔥 LAUNCH: $19.50 with code EARLYBIRD
VRAM: ~4.8GB • NVIDIA only
Get Nemotron — $39Professional
Kyutai STT 2.6B • English • 5.8GB VRAM
one-time
🔥 LAUNCH: $29.50 with code EARLYBIRD
VRAM: ~5.8GB • NVIDIA only • ~2s delay
Get Pro Kyutai 2.6B — $59Performance
Kyutai STT 1B • EN + FR • 2.7GB VRAM
one-time
🔥 LAUNCH: $19.50 with code EARLYBIRD
VRAM: ~2.7GB • NVIDIA only • <80ms
Get Kyutai 1B — $39Flagship
Voxtral 4B • Mistral AI • 14GB VRAM
one-time
🔥 LAUNCH: $29.50 with code EARLYBIRD
VRAM: ~14GB • NVIDIA only • Docker required
Get Pro Docker — $59No Docker
Voxtral F16 4B • Mistral AI • 8.5GB VRAM
one-time
🔥 LAUNCH: $29.50 with code EARLYBIRD
VRAM: ~8.5GB • NVIDIA only • No Docker
Get Pro Native — $59Just want to try it? VoxBar Whisper is free — runs on any hardware, no GPU needed.
Download Free — VoxBar Whisper ($0)| Free | GLM | Nemotron | Kyutai 2.6B | Kyutai 1B | Pro | |
|---|---|---|---|---|---|---|
| Engine | VoxBar Faster-Whisper (2GB VRAM) | VoxBar Canary Qwen 2.5B (4GB VRAM) | VoxBar Nemotron 0.6B (2GB VRAM) | VoxBar Kyutai 2.6B (6GB VRAM) | VoxBar Kyutai 1B (2.7GB VRAM) | VoxBar Voxtral 4B (14GB VRAM) |
| Speed | ⚡ Fast | ⚡ Fast | ⚡⚡ Very Fast | ⚡ Delayed (~4s) | ⚡⚡⚡ Instant | ⚡⚡⚡ Instant |
| Accuracy | Good | Great | Great | Excellent | Very Good | S-Tier |
| Languages | English | English | English | EN + FR | EN + FR | 13+ |
| GPU | CPU only | NVIDIA | NVIDIA | NVIDIA | NVIDIA | NVIDIA/AMD/Apple |
| Docker | — | — | — | — | — | Win only |
| Platform | Windows | Windows | Windows | Windows | Windows | Win + Mac 🍎 |
| Price | Free | $39 | $39 | $59 | $39 | $59 |
All plans include Overlay Mode • Standard View • Voice Commands • Lifetime License • Free Updates
A floating voice bar that disappears into your workflow. Dictate emails in Outlook, narrate code in VS Code, write reports in Google Docs — blending seamlessly into any background with adaptive themes and Glass Mode.
On dark backgrounds, the bar is completely invisible. Just words appearing out of thin air. Adjust opacity from 25% to 100%.
Drag the bar to any edge of your screen. It remembers its position between sessions. Resize it to fit any space.
Say "delete" to remove the last word. Dictate punctuation naturally: "comma", "full stop", "new paragraph", "open bracket".
✨ Live Demo — Watch Overlay Mode in Action
The quarterly report shows a 23% increase in user engagement across all platforms. This growth is primarily driven by improvements in our mobile experience and the launch of three new features.
Revenue increased by $2.4M compared to the previous quarter, with customer satisfaction reaching an all-time high of 4.8/5.0 across every region.
Key highlights include: reduced churn by 12%, increased trial-to-paid conversion rate by 18%, and expanded into 3 new markets. The engineering team shipped 47 features with a 99.9% uptime...
|
🪟
Multi-Monitor
Open a overlay bar on each display. They work independently.
🧲
Drag Like a Toy
Grab it, move it, resize it. It snaps to edges and remembers position.
🎯
Click to Listen
Listens only when you tell it to. Hit the button, speak, and it auto-stops when you're done.
The full-featured window for when you need to see every word. Dictate documents, emails, clinical notes, and code — in 13 languages, more coming soon.
Chapter 12: The Morning Light
The morning light filtered through the curtains, casting long golden shadows across the wooden floor. She sat at her desk, coffee growing cold beside her, fingers hovering above the keyboard.
But today was different. Today she didn't need to type.
|
Composing reply...
Dictating notes...
VoxBar™
📝 Writers & Authors — Dictate directly into your documents
ENEnglish
FRFrench
ZHChinese
ARArabic
DEGerman
JAJapanese
ESSpanish
HIHindi
ITItalian
KOKorean
PTPortuguese
RURussian
NLDutch
More languages coming soon.
You've got the GPU — put it to work. Dictate scripts, take notes, and transcribe offline.
Transcribe lectures offline. Review notes instantly. No monthly fees.
Write novels by speaking. No subscription limits on word count.
Dictate sensitive client notes without violating NDAs or HIPAA.
Each tier has different hardware needs. Pick the one that fits your setup.
Lightweight engines — runs on almost anything
Canary Qwen 2.5B engine — best balance of speed + accuracy
Voxtral 4B engine — maximum intelligence
VoxBar™ Pro requires Docker Desktop and an NVIDIA GPU.
Same Voxtral 4B engine — native Apple Silicon
Metal GPU acceleration • No Docker • One-click install
All tiers include lifetime license + future updates
Some Windows users may see a SmartScreen prompt during installation. VoxBar is an independently developed utility — we don't gather or share your data, which means Microsoft doesn't "know" us yet. This is completely normal for independent software.
Quick Installation Steps (Windows):
Mac Users:
VoxBar Pro on Mac runs natively — no Docker, no Gatekeeper issues. Just double-click the launcher and the engine builds itself automatically.
The Story
Real‑time, multilingual speech‑to‑text isn't science fiction anymore — it's already a reality. Meet VoxBar™: the next generation of human‑to‑machine communication, bridging the gap between your voice and your computer. If you're into coding, creative work, or anything that demands fast output, chances are your hardware is already ready. Imagine all‑day productivity boosts you never believed were possible. That's what happens when you remove the keyboard from the equation.
And here's the part most people miss: the biggest barrier to AI adoption isn't the technology — it's trust. People are reluctant to speak freely when big tech is listening. Client contracts, medical records, business strategy, personal reflections — none of it belongs on someone else's server. No cloud listening in. No big tech harvesting your words. VoxBar™ was built from day one with a simple rule: we don't gather your data. Period.
A developer built it for himself. It was too good not to share. Voice‑first computing isn't a novelty — it's the future of how we work. And the future starts on your machine.
Read the Full Story →