See it in action
Meet WritHer.
What you get
Everything runs
on your machine.
No subscriptions. No internet after setup. Your voice data never leaves your computer.
Universal Dictation
Hold a key, speak, release. Text is pasted into whichever window has focus — editors,
browsers, chat apps, anything.
Voice AI Assistant
Natural language commands parsed by a local LLM via Ollama. Save notes, schedule
appointments, set reminders — all by voice.
100% Offline
Speech recognition via faster-whisper, intent parsing via Ollama, data stored in local
SQLite. Zero cloud dependency.
Smart Scheduling
Say "meeting next Monday at 3pm" and the LLM converts relative times to absolute
datetimes automatically.
Animated Widget
A minimal pill-shaped overlay with expressive "Pandora Blackboard" eyes that react to
state: listening, thinking, happy, error.
Multi-language
Ships with English and Italian. Easy to add more via the locales.py string table — no
recompilation needed.
Two modes, one tray icon
Hold a key.
Speak. Done.
Hold
AltGr
Dictation
Transcribes your voice and pastes the text directly into whichever app has focus.
Clipboard is saved and restored automatically.
Write a quick email reply
Type in VS Code, Notion, Teams...
Any text field, any app
Hold
Ctrl
R
Assistant
Natural-language commands powered by a local Ollama LLM.
Create structured data from speech without touching the keyboard.
Remind me to call Marco in one hour
Shopping list: bread, eggs, butter
Dentist appointment tomorrow at 3pm
Honest comparison
How WritHer stacks up.
Looking for a free, open-source Wispr Flow alternative or a private replacement for Dragon NaturallySpeaking? WritHer is the only tool that combines universal dictation with a local LLM assistant — entirely on your device.
| Feature | Writher | Wispr Flow | Dragon | Win. Speech | Whisper Desktop |
|---|---|---|---|---|---|
| 100% offline / local | ✓ | ✗ cloud | ✓ | ✓ | ✓ |
| Free & open source | ✓ | ✗ ~$12/mo | ✗ $150–500 | ✓ | ✓ |
| Works in any app | ✓ | ✓ | ✓ | ⚠ limited | ✓ |
| Voice AI assistant | ✓ | ✗ | ✗ | ✗ | ✗ |
| Notes & reminders by voice | ✓ | ✗ | ✗ | ✗ | ✗ |
| Local LLM (Ollama) | ✓ | ✗ | ✗ | ✗ | ✗ |
| No account / API key | ✓ | ✗ | ✓ | ✓ | ✓ |
| Data never leaves device | ✓ | ✗ | ✓ | ✓ | ✓ |
| Multi-language | ✓ | ✓ | ✓ | ⚠ limited | ✓ |
| Animated overlay widget | ✓ | ✗ | ✗ | ✗ | ✗ |
Built with
Open source,
end to end.
🎤 faster-whisper — speech recognition
🧠 Ollama — local LLM inference
🐍 Python 3.11+
🗄️ SQLite — local database
🖼️ Tkinter — widget & overlay
🔔 winotify — Windows toast notifications
⌨️ pynput — global hotkeys
📋 Win32 API — clipboard injection
Your voice.
Your data.
Your rules.
Free. Open source. MIT licensed.
No cloud. No subscriptions. No telemetry. Ever.
💛 Contributors Welcome
WritHer is built by the community, for the community. Are you a Python developer, UI enthusiast, or translator? We are actively looking for collaborators to help shape the future of local AI dictation.
Explore Open Issues