Whisper ASR

🎙️ Whisper ASR — Audio & Video Transcriber 🎧 Ultra‑fast, on‑device transcription — up to 20× faster than real‑time 22 minutes transcribed in 58 seconds using the fastest GGML Whisper model (~20× real‑time) — GPU-enabled Experience fast, private, and GPU‑accelerated speech‑to‑text with a full Whisper‑powered transcription suite built for creators, developers, and power users. This app brings official OpenAI Whisper models directly to your desktop with seamless Hugging Face Hub integration, resume‑supported downloads, and a clean GGML model browser. Run everything locally — no cloud, no data sharing, no latency. Choose from English‑only or multilingual models, including official FP16 Whisper releases and high‑performance GGML quantized variants. A built‑in GGML Model Server lets you route, manage, and benchmark more than 30 Whisper models across tiny, base, small, medium, and large tiers. Unlock maximum performance with NVIDIA GPU acceleration powered by CUDA. Monitor real‑time inference speed, decoding throughput, and GPU utilization with the integrated performance dashboard. Whether you're transcribing long‑form audio, building AI workflows, or deploying Whisper in production, this app delivers a fast, secure, and professional‑grade transcription experience. (Benchmark hardware: RTX 4050 (6GB VRAM) + AMD Ryzen 9 — performance varies by hardware) 🔐 Licensing & Activation Whisper ASR is free to install from the Microsoft Store. To unlock full access, a license key is required. - License keys are available via Gumroad for a one-time purchase of $59 - Note: Whisper ASR license keys are not purchasable directly through the Microsoft Store 🗝️ Purchase license key: https://productgenius.gumroad.com/l/whisper-asr 🎓 What’s Included: • ✅ One-time purchase — lifetime access to all future updates • 💸 No subscriptions or monthly fees • 💡 A transparent, cost‑effective alternative to expensive cloud‑based transcription services ⭐ Core Features & Highlights • ♾️ Unlimited audio & video transcriptions • 🌍 Transcribe speech in 100 languages • 🈂️ Translate to English • ⚡ Local GPU acceleration • 🚫 Zero per‑minute cost • 🔒 No data leaves the device • 🔌 No API dependency • 🔓 No usage caps 🤗 Official OpenAI Whisper Models • Hugging Face Hub — GGML Model Browser • Direct Download (resume-supported) • Full library, 33 GGML Whisper models • 11 official FP16 • 22 optimized quantized variants • Tiny → Base → Small → Medium → Large (v1, v2, v3, turbo) 🔒 Private, Local AI/ML Transcription • 100% on‑device processing • English‑only and multilingual models • Official FP16 Whisper models + GGML quantized variants • Built on the OpenAI Transformer architecture ✳️ GPU Acceleration • NVIDIA GPU‑powered high‑performance inference • Optimized for HPC/AI workloads • Whisper.cpp CUDA runtime for maximum throughput ⚡ Real-time GPU Performance Monitor • Live decoding + transcription telemetry • GPU utilization, inference speed, and decoding performance • Built‑in diagnostics for CUDA‑accelerated ASR 🎙️ Audio & Video Transcription • 🎧 Transcribe MP3, WAV, FLAC, OGG (Vorbis), MP4, MOV, WEBM, and MKV • ▶️ YouTube support — download videos as MP4 and transcribe instantly • 🎤 Built‑in Recorder — capture meetings, lectures, or conversations using laptop or external microphones 📤 Transcript Tools + Export Options & Translation • 📋 Copy to Clipboard for quick use in docs or productivity apps • 🔎 Raw SRT viewer with timestamp navigation and highlighted keyword search (original + translated SRT) • 📥 Export in 17+ formats: TXT, VTT, SRT, HTML, LRC, CSV, DOCX, PDF, JSON, JSON‑FULL, SBV, TTML/DFXP, ASS/SSA, RTTM, CTM • 🔗 HTML (Standalone Viewer) with dark/light themes and instant re‑export (JSON, VTT, CSV, TXT). Share anywhere — drop it in an email and your collaborators can open it immediately • 🤝 Share transcripts via email for review or collaboration • 🌍 Translate transcriptions into 136 Languages (internet required). Note: Online translation may be temporarily unavailable during periods of limited server availability. 🎧 Playback, Review & Sync Tools • 🎼 Audio‑Synced LRC Transcript Player — auto‑synced timeline for audio & video • Upload existing audio/video with .lrc for synchronized timeline playback • 📦 Supports files & transcripts up to 200MB for stable, fast processing 🖥️ GGML Model Server • Route, manage, and brand 33 Whisper transcriber models (11 official + 22 quantized) • Unified runtime across tiny → base → small → medium → large models (v1, v2, v3, turbo) • Ideal for multi‑model workflows and benchmarking ⚙️ Performance & System Modes • 🖥️ GPU-accelerated for maximum speed or CPU‑Only for benchmarking • 🖥️ Run anywhere — from laptops to high‑end NVIDIA GPUs • 📦 Supports up to 1GB per transcription, enough for 10+ hours of audio. ⚙️ System Guidance • 8–12 GB RAM: Best for short recordings and everyday transcription • 16–32 GB RAM: Ideal for creators and multi‑hour audio • 64–128 GB RAM: Optimized for large models and high‑volume workloads • This app supports audio & video files up to 1 GB, typically equal to 10+ hours of compressed formats like MP3, M4A(AAC/ALAC), or Opus. • Performance and actual capacity vary based on your VRAM, RAM, available disk space, and the model size you’re using. ⭐ System Requirements Windows 10/11 🆕 Supported NVIDIA GPU RTX Generations RTX Generation | Architecture | Compute Capability ---------------|---------------|-------------------- RTX 20‑series | Turing | 7.5 RTX 30‑series | Ampere | 8.6 RTX 40‑series | Ada Lovelace | 8.9 NVIDIA TITAN Series (Various Architectures) NVIDIA Quadro & RTX A‑Series (Workstation GPUs) ⚠️ Compatibility Notice ❌ RTX 50‑series | Blackwell | 12.0 - Currently not supported upstream. Whisper ASR supports NVIDIA GPUs with modern CUDA drivers. Older GPUs may not be compatible with the CUDA backend. ❌ Not supported for CUDA acceleration: • Fermi‑generation GPUs (e.g., GeForce 300M / 400M / 500M) • Maxwell GPUs (e.g., GTX 9xx, 940M) • GTX 10‑series and GTX 16‑series (Pascal) • GPUs are limited to legacy drivers that do not include modern CUDA features These older GPUs may fail to load the CUDA backend due to missing driver functions. 🎯 Who This Is For • Creators & podcasters • Students & educators • Professionals & teams • Developers & AI builders • Privacy‑focused users • Accessibility & support roles 🌐 Translate into 136 languages Major families supported: • European: English, French, German, Spanish, Italian, Portuguese, Dutch, Swedish, Polish, Romanian, Czech, Greek • Asian: Chinese (Simplified/Traditional), Japanese, Korean, Hindi, Bengali, Tamil, Telugu, Thai, Vietnamese • Middle Eastern: Arabic, Hebrew, Persian, Kurdish (Sorani/Kurmanji), Turkish • African: Swahili, Yoruba, Zulu, Hausa, Igbo, Amharic • Pacific & Indigenous: Maori, Samoan, Hawaiian, Aymara, Quechua • Plus 100+ additional languages Note: Online translation may be temporarily unavailable during periods of limited server availability. 🔗 Visit @Generative Lab: https://genlab.i-mapz.com

Utilities · Windows · Free

Screenshots

Download

• Microsoft Store