Writher – offline voice assistant for Windows (Whisper and Ollama)
Fully-local Whisper+Ollama voice stack; real Windows app with emoji widget.

Full-stack browser voice AI (WebLLM, Whisper.cpp, VITS) running 100% locally and client-side.
AI enthusiasts and developers interested in local, privacy-preserving voice assistants running on WebGPU.
Ollama · LM Studio · Jan.ai
I've been experimenting with pushing local AI fully into the browser via Web Assembly and WebGPU, and finally have a semblance of a working platform here! It's still a bit of a PoC but hell, it works.
You can create assistants and specify: - Wake word - Language model - Voice
I published a little post on my site with some technical details as well if anyone is interested: https://shaneduffy.io/blog/i-built-a-voice-assistant-that-ru...
Fully-local Whisper+Ollama voice stack; real Windows app with emoji widget.
One Rust binary does what Electron apps and Python scripts couldn't for Linux dictation.
Hand-tuned SSE particle engine from 2002 assembly, now runs in your browser via WASM.
Full-featured Android voice assistant, but tied to OpenClaw ecosystem adoption.
The repo solves a real Wayland pain with a clever two-part design: a lightweight out-of-sandbox trigger script wakes a sandboxed app via DBus so you can have a global hotkey without breaking security. It runs whisper.cpp locally and shows an overlay to paste transcripts system-wide — very useful for offline, privacy-first typing — but the Flatpak/runtime, AVX2 needs, and English-only support limit its reach.
Offline voice typing for Linux—finally fills a gap Windows and macOS solved years ago.