Hey everyone! I wanted to share a small tool I’ve been building called WritHer.(Free and open source alternative to Wispr Flow)
The idea is simple: it lives in your system tray and gives you two things.
Hold AltGr anywhere (any app, any text field) and just speak. It transcribes your voice with Whisper and pastes the text right where your cursor is. No clicking, no switching apps.
Hold Ctrl+R and you get a voice assistant that understands natural language. You can say things like “remind me to call Marco in one hour” or “appointment with the dentist tomorrow at 3pm” and it handles the rest. Notes, to-do lists, shopping lists, reminders with toast notifications, all stored locally in SQLite.
The part I’m most proud of: everything runs 100% offline. Speech recognition via faster-whisper, intent parsing via Ollama, no cloud, no API keys, no telemetry. Once you download the models it works with no internet at all.
There’s also a little animated floating widget with eyes that react to what it’s doing (listening, thinking, error…) which is silly but I kind of love it.
It’s Python, MIT license, Windows 10/11 only for now.
GitHub: https://github.com/benmaster82/writher
Would love feedback, especially from anyone who uses voice input regularly. Still early days but it works well for my daily workflow!