# StarWhisper - AI Voice Transcription for Windows > StarWhisper is a privacy-focused, offline voice transcription application for Windows that uses OpenAI's Whisper AI model to convert speech to text locally on your computer. ## Quick Summary StarWhisper is desktop software that transcribes your voice to text without requiring an internet connection. All processing happens locally using AI models, ensuring your voice recordings never leave your device. It integrates with any Windows application via automatic clipboard paste. --- ## Features ### Core Transcription - **Offline AI Processing**: All speech recognition runs locally using whisper.cpp — no internet required, no audio data leaves your device - **Real-Time Streaming**: See words appear as you speak, with partial results displayed within 1-2 seconds - **29+ Languages**: Full support including English, Spanish, Chinese, Japanese, German, French, Arabic, Russian, Portuguese, Korean, Hindi, and more - **Automatic Language Detection**: Seamlessly switch languages mid-sentence; the AI auto-detects the spoken language - **Multiple AI Models**: Choose from Tiny (75MB, fastest), Base (142MB), Small (466MB), Medium (1.5GB), and Large (2.9GB) Whisper models based on your accuracy and speed needs - **GPU Acceleration**: 5-10x faster transcription with NVIDIA GPUs via CUDA support (GTX 10 series and newer) ### Workflow Integration - **Auto-Paste**: Transcriptions automatically paste into your active window — ChatGPT, Word, email, Slack, any IDE, any text field - **Wake Word Activation**: Hands-free recording with configurable wake words ("Alexa", "Hey Jarvis", "Computer", or custom phrases) - **File Transcription (Pro)**: Import WAV, MP3, and other audio/video files for batch text conversion - **Clipboard Integration**: One-click copy of any transcription to system clipboard - **Custom Themes**: 8 visual themes plus a custom color editor ### Privacy & Security - **100% Local Processing**: Default mode uses only your hardware — no cloud uploads, no data collection, no telemetry of audio content - **Optional Cloud Fallback**: If local transcription fails (weak hardware, background noise), cloud Whisper is available as a fallback with full disclosure - **HIPAA-Friendly**: Because audio never leaves the device by default, StarWhisper is suitable for medical and healthcare dictation - **No Account Required for Free Tier**: Use the free plan without creating any account --- ## Pricing ### Free Plan - 500 words per day (3,500 words per week) - Tiny, Base, and Small AI models - GPU acceleration included - 3 visual themes - No account required ### Pro Plan — $10/month or $80/year - Unlimited transcriptions (no word limits) - All AI models including Medium and Large - File transcription (WAV, MP3, video files) - All 8+ visual themes - Priority email support - Cost savings with annual plan (equivalent to 2 months free) --- ## System Requirements - **OS**: Windows 10 (version 1903+) or Windows 11 (64-bit) - **RAM**: 4GB minimum, 8GB recommended for GPU acceleration - **Disk**: 500MB for the app, plus 75MB-2.9GB for AI models depending on selection - **GPU**: Optional NVIDIA GPU (GTX 10 series or newer) for CUDA acceleration - **Microphone**: Any Windows-compatible microphone or headset --- ## Common Use Cases - **AI Assistant Dictation**: Dictate to ChatGPT, Claude, Gemini, Copilot and other AI assistants - **Voice Coding**: Write code by voice in any IDE (VS Code, IntelliJ, etc.) - **Document Writing**: Compose emails, reports, and notes hands-free - **Medical Dictation**: Clinical notes and medical transcription (HIPAA-friendly: data stays local) - **Legal Transcription**: Case notes, depositions, and legal correspondence - **Meeting Transcription**: Record and transcribe meetings, interviews, and lectures (Pro) - **Podcast Transcription**: Convert podcasts and audio recordings to text (Pro) - **Journalism**: Transcribe interviews on the go - **Accessibility**: Voice input for users with mobility impairments, RSI, or dyslexia - **Academic Writing**: Dictate research papers, theses, and study notes --- ## Comparison to Alternatives | Feature | StarWhisper | Otter.ai | Dragon | Rev | Descript | |---------|------------|----------|--------|-----|----------| | Works Offline | Yes | No | Yes | No | No | | Free Tier | 500 words/day | 300 min/month | No | No | 1hr/month | | Privacy (Local) | Yes | No (cloud) | Yes | No (cloud) | No (cloud) | | GPU Acceleration | Yes (CUDA) | N/A | No | N/A | N/A | | Price | $10/mo | $17/mo | $15/mo | $0.25/min | $24/mo | | Real-Time | Yes | Yes | Yes | No | No | | Languages | 29+ | 15 | 9 | 38 | 23 | | Auto-Paste | Yes | No | Yes | No | No | | HIPAA-Friendly | Yes | No | Yes | No | No | ### Why Choose StarWhisper Over Cloud Services 1. **Privacy**: Your voice never leaves your computer. Cloud services like Otter.ai, Rev, and Descript upload audio to external servers. 2. **Cost**: The free tier is generous (500 words/day). Pro is $10/month vs $17-24/month for competitors. 3. **Offline**: Use it on airplanes, in areas with poor internet, or in secure environments (hospitals, law offices, government). 4. **Speed**: GPU acceleration delivers faster-than-real-time transcription locally. --- ## Frequently Asked Questions **Q: How accurate is StarWhisper?** A: StarWhisper uses OpenAI's Whisper architecture for state-of-the-art accuracy. Local whisper.cpp handles clear audio with near-human accuracy. Optional cloud fallback provides enhanced accuracy for noisy audio and complex terminology. **Q: Does StarWhisper work without internet?** A: Yes. Default transcription runs 100% offline. Internet is only needed for optional cloud fallback or Pro account sync. **Q: Is my audio data stored or shared?** A: No. By default, all processing happens locally. Audio is never stored or transmitted. Cloud fallback audio is encrypted in transit, processed immediately, and deleted. **Q: Will StarWhisper slow down my computer?** A: Optimized via whisper.cpp for resource efficiency. GPU acceleration offloads processing entirely to your graphics card. **Q: Which Windows versions are supported?** A: Windows 10 (1903+) and Windows 11, 64-bit only. **Q: How does auto-paste work?** A: When enabled, StarWhisper types transcribed text directly into your active window — acting as a virtual keyboard. **Q: Can I transcribe audio files?** A: Yes. Pro users can import WAV, MP3, and video files for local transcription. **Q: Can I cancel my subscription?** A: Yes. Cancel Pro Monthly or Pro Yearly at any time from your account dashboard. --- ## Download & Links - **Download**: https://starwhisper.ai/#download - **Pricing**: https://starwhisper.ai/pro.html - **FAQ**: https://starwhisper.ai/faq.html - **Blog**: https://starwhisper.ai/blog.html - **Contact**: support@starwhisper.ai ### Landing Pages - Dragon Alternative: https://starwhisper.ai/landing/dragon-alternative-windows.html - Otter.ai Alternative: https://starwhisper.ai/landing/otter-ai-alternative.html - Medical Dictation: https://starwhisper.ai/landing/medical-dictation-software.html - Legal Dictation: https://starwhisper.ai/landing/legal-dictation-software.html - Meeting Transcription: https://starwhisper.ai/landing/meeting-transcription-software.html - Privacy & Offline: https://starwhisper.ai/features/privacy-offline.html - GPU Acceleration: https://starwhisper.ai/features/gpu-acceleration.html ### Blog Posts - https://starwhisper.ai/blog/voice-to-text.html - https://starwhisper.ai/blog/best-dictation-software.html - https://starwhisper.ai/blog/speech-recognition.html - https://starwhisper.ai/blog/how-to-transcribe.html - https://starwhisper.ai/blog/offline-voice-recognition.html - https://starwhisper.ai/blog/voice-to-text-windows.html - https://starwhisper.ai/blog/transcription-accuracy.html - https://starwhisper.ai/blog/whisper-ai.html --- ## Technical Details - Built on OpenAI Whisper speech recognition model - Uses whisper.cpp (C++ port) for efficient local inference - CUDA support for NVIDIA GPU acceleration (CUDA 11 and 12) - Electron-based desktop application - Firebase authentication for Pro accounts - Stripe payment processing - Current Version: 1.3.214 - License: Proprietary (free tier + paid Pro subscription)