MeshUtility
A tray-resident widget that does two things anywhere on your system: dictate speech into whatever field has focus, and rewrite text you have selected. Built with Tauri, Rust and React — and it works with or without MeshConsole.
MeshVoice — dictate into anything
Hold the hotkey, speak, release. What you said is transcribed and typed into whatever window has focus — a terminal, an editor, a browser field. Push-to-talk or toggle, and the hotkey is yours to choose.
- Local transcription with Parakeet V3 — offline, nothing leaves the machine
- A cloud engine when you want the speed instead — one toggle apart
- Custom dictionary, replayable history, and session stats (words, minutes, WPM)
Open source, and it stays out of the way
MeshUtility sits in the system tray and costs nothing to leave running. Voice history, your dictionary and your provider settings are stored locally, and the whole thing is Apache-2.0 on GitHub — read it, build it, send a pull request.
- Your own key, your own provider: OpenAI, Anthropic, Gemini, Groq, Mistral or local Ollama
- Pick how much rewriting you want, from concise to expansive
- Lives in the tray; Apache-2.0, and the source is on GitHub
Rewrite what you selected, wherever you selected it
Select text anywhere and hit the overlay hotkey. MeshPrompt sharpens a vague prompt, tightens prose, fixes grammar or translates it — then copies the result or replaces the original in place. Choose how far it should go, from concise to expansive.
It runs on your own key and your own provider: OpenAI, Anthropic, Google Gemini, Groq, Mistral, or a local Ollama model if you would rather nothing left the machine at all.