AI Speech-to-Text Dictation. Instantly.
The Fastest, Privacy-First Speech-to-Text Desktop App — Built natively in Rust.
Perpetual license.
Sign up for the waitlist to secure this discounted rate when we launch.
$ minerva start --mode=email
> Recording initialized (Ctrl+Shift+Space)
Transcribing... "Hey team just wanted to check on the Q3 roadmap."
> Processing via AI Mode [Email]
I just wanted to touch base and check on our progress regarding the Q3 roadmap.
Best regards,
# Typed directly at your cursor. In seconds.
Try different modes
Blazing Fast
Built entirely in Rust for native performance. Near-zero latency recording and transcription that keeps up with your thoughts.
Ultra Lightweight
Minimal memory footprint means Minerva runs quietly in the background without slowing down your computer.
Privacy First
Your audio recordings stay on your machine. Only transcriptions are sent to your chosen AI provider securely.
One Hotkey Away
Press your custom hotkey, speak your thoughts, and watch text appear at your cursor. No context switching.
Core Features
Advanced AI Transcription, simplified.
Instant Voice Dictation
Push-to-talk recording with customizable global hotkeys. Watch text appear right at your cursor in any application — email, documents, chat, or code editors.
AI-Powered Text Transformation
Minerva doesn't just transcribe — it transforms. Switch between Basic (raw text), Email (professional formatting), Casual (conversational tone), or create custom prompts.
Bring Your Own AI
Choose your preferred transcription and LLM providers. Built-in support for Groq Cloud, Fireworks AI, and Cerebras, plus extensible TOML configs for OpenAI or local models via Ollama.
History & Dashboard
Browse past transcriptions, check API costs, monitor word counts, and easily manage your audio files with configurable automatic cleanup limits.
Local Speech-to-Text Models
Support for both inference from AI providers (Groq, OpenAI, etc.) as well as local models! Available models include Whisper (Multilingual, 100+ languages), Parakeet V3 (25 European Languages), Moonshine Base (English), and SenseVoice (Asian languages).
Beautiful Native UI
Whether you prefer the sleekness of dark mode or the clarity of light mode, Minerva has you covered. Full theming support explicitly built to match your desktop OS preferences seamlessly.
Speech-to-Text Uses
Built for everyday professional workflows.
Business Professionals
Dictate emails, meeting notes, and reports. Email mode formats your speech into polished correspondence automatically.
Writers & Creators
Capture ideas as fast as you can speak them. Use casual mode to generate conversational drafts instantly.
Developers
Add comments, write documentation, and respond to messages without leaving your IDE.
Students & Researchers
Transcribe lectures and interviews. Keep detailed notes by voice.
Accessibility
Voice-first input for users with mobility or typing challenges.
Why Minerva stands out
| Feature | Minerva | Other STT Apps |
|---|---|---|
| Built in Rust | ✅ | ❌ |
| Custom LLM Modes | ✅ | ❌ |
| Multiple Providers | ✅ | ❌ |
| System Tray Integration | ✅ | Varies |
| Memory Usage | ~50MB | 200MB+ |
| Startup Time | Instant | Varies |
| Offline Recordings | ✅ | ❌ |
| Privacy-First | ✅ | Varies |
| Auto Updates | ✅ | Varies |
Technical Highlights
- System
- Single Instance, Global Hotkeys OS Keyring Storage
- Audio Engine
- WAV Recording, PulseAudio/PipeWire
- Platforms
- Windows x86_64, Linux x86_64/aarch64
- Memory Safety
- Zero-Cost Abstractions, No data races
Get on the Waitlist
$5 Perpetual License (Regularly $15)
Secure this introductory pricing by joining the waitlist today.
Minerva will be a paid application. Join the waitlist now to lock in the introductory $5 perpetual license pricing when we launch. Join the professionals and developers who dictate flawlessly without context switching.