The latest articles on DEV Community by BrethofAI (@brethofai). https://dev.to/brethofai https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3878857%2Fd0b3f37e-54d4-4266-be08-06d204e8bb1a.jpg https://dev.to/brethofai en BrethofAI Thu, 16 Apr 2026 07:32:31 +0000 https://dev.to/brethofai/i-built-a-49-voice-to-text-app-that-never-touches-the-cloud-4b4g https://dev.to/brethofai/i-built-a-49-voice-to-text-app-that-never-touches-the-cloud-4b4g <h1> Why I Built Brethof Voice Pro </h1> <p>I got tired of two things:</p> <ol> <li>Voice-to-text tools that only work well in English</li> <li>Every good STT solution either costs $700 (Dragon) or sends your voice to the cloud monthly</li> </ol> <p>So I built something different.</p> <h2> What It Does </h2> <p>Brethof Voice Pro is a desktop app (Windows + Linux) that converts speech to text using AI — entirely on your device. Press Ctrl+D, speak, text appears where your cursor is.</p> <p><strong>Zero cloud calls during transcription.</strong> Models download once on first launch, then it works fully offline.</p> <h2> The Tech Stack </h2> <ul> <li> <strong>Qwen3-ASR engine</strong> — 1.84% word error rate across 10 languages (arXiv 2601.21337)</li> <li> <strong>GGUF models via llama.cpp</strong> — 6 quantization tiers from 1 GB to 3.2 GB</li> <li> <strong>Vulkan GPU acceleration</strong> — works on NVIDIA, AMD, Intel, and CPU-only</li> <li> <strong>DeepFilter</strong> noise reduction</li> <li> <strong>36 languages</strong> with auto-detection</li> </ul> <h2> Why It Matters </h2> <p>If you speak Thai, Polish, Arabic, Vietnamese, Korean, or dozens of other languages — there has been <strong>no good voice-to-text option</strong> for you. Dragon doesn't support most languages. Google's API charges per-minute and requires internet. Whisper's accuracy on non-English languages is mediocre.</p> <p>Qwen3-ASR changes this. State-of-the-art accuracy across 36 languages, running locally on consumer hardware.</p> <h2> Pricing </h2> <p>$49 one-time. Perpetual license. No subscription.</p> <p>Compare: Dragon $699. Otter.ai $100-240/year. Google STT per-minute cloud pricing.</p> <p>14-day free trial, no credit card: <a href="https://brethof.com" rel="noopener noreferrer">brethof.com</a></p> <p>Happy to answer questions about the architecture, GGUF quantization, or Vulkan vs CUDA tradeoffs.<br> <a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fdla86uwsr756ld39t3ph.png" class="article-body-image-wrapper"><img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Farticles%2Fdla86uwsr756ld39t3ph.png" alt=" " width="800" height="264"></a></p> ai privacy linux productivity