Voice transcription for everyone. No GPU. No cost. No catch.
Powered by Faster Whisper — OpenAI's battle-tested speech model optimized with CTranslate2, runs on any hardware including CPU-only machines.
Lightweight chunk-based processing. Runs on any machine — even without a dedicated GPU.
Captures audio at 16kHz using sounddevice. Low resource footprint from the start.
Collects a few seconds of audio. Silence is detected and skipped to avoid wasting processing power.
The optimized Faster Whisper model processes each chunk on your CPU or GPU. No cloud, no internet, everything stays on your machine.
Clean text is appended. Each chunk is independent — no growing state, no degradation over time.
VoxBar Free is a great starting point. When you're ready for more, upgrade to a paid tier:
| Requirement | Minimum |
|---|---|
| GPU | Not required (CPU works) |
| RAM | 4GB |
| Disk | ~200MB |
| OS | Windows 10/11 |
| Docker | Not required |
No sign-up. No payment. No catch. Just download and start speaking.
Download VoxBar Free