4.4 KiB
Hyprvoice
Voice‑powered typing for Wayland/Hyprland desktops — press a global shortcut, speak, and watch words appear in whatever window you’re focused on.
Why does this hyprvoice exist?
Typing is slow, repetitive strain injuries are real, and proprietary speech‑to‑text solutions leak your data. This project aims to give Wayland users a fast, privacy‑respecting alternative that works entirely on their own hardware or any cloud ASR they trust.
Key Features
| Category | Highlights |
|---|---|
| Hot‑key daemon | Background service triggered by a global shortcut (via hyprctl dispatch exec … or other compositor hooks). |
| Live audio capture | Low‑latency PipeWire input with optional VAD/noise gate. |
| Pluggable ASR | 🔌 Local Whisper (via whisper.cpp) or cloud APIs (OpenAI, Deepgram, AssemblyAI). Select at runtime. |
| Optional LLM polish | Feed raw transcript into a local Ollama model or GPT‑4o to add punctuation, change tone, etc. |
| Text injection | wtype/ydotool sends keystrokes to the focused surface; clipboard fallback for stubborn clients. |
| Zero‑friction config | Single ~/.config/hyprvoice/config.toml with sensible defaults and live reload on SIGHUP. |
| Cross‑platform binary | Static Go build; no Python environment required. |
Quick Start (Arch / Hyprland)
# 1. Install from AUR (source build)
yay -S hyprvoice
# or pre‑built binary
yay -S hyprvoice-bin
# 2. Run the interactive setup helper
hyprvoice-install run --backend whispercpp
# 3. Enable the user service
systemctl --user enable --now hyprvoice.service
# 4. Add a key binding in Hyprland conf
bind = $mod, R, exec, hyprvoice toggle
Configuration file (~/.config/hyprvoice/config.toml)
[asr]
backend = "whispercpp" # whispercpp | openai
model = "medium.en.bin" # used if backend = whispercpp
api_key = "" # used if backend = openai
[vad]
threshold = -45 # dBFS
hang_ms = 300
[inject]
method = "wtype" # wtype | ydotool | clipboard
[keybind]
# If you don't use Hyprland, set an XDG desktop accelerator here
shortcut = "CTRL+ALT+SPACE"
All options are documented in the sample config generated by --init.
Architecture Overview
┌──────────┐ PCM ┌──────────┐ chunks ┌──────────┐ text ┌──────────┐
│ Hot‑key │────────▶│ Audio │────────▶│ ASR │──────▶│ Inject │
│ Listener │ │ Capture │ │ Adapter │ │ (wtype) │
└──────────┘ └──────────┘ └──────────┘ └──────────┘
▲ │ │
│ stats / errors │ ▼
└──────────────┬────────┘ ┌──────────┐
│ │ Desktop │
└──────────────────▶ UI │
└──────────┘
Each box is a Go package inside internal/ so you can swap implementations without touching public APIs.
Building from Source
git clone https://github.com/leonardotrapani/hyprvoice.git
cd hyprvoice
# Compile
CGO_ENABLED=1 go build -o hyprvoice ./cmd/hyprvoice
# Run tests
go test ./...
Requires Go 1.22+, a C compiler, and pkg-config with PipeWire headers.
Contributions welcome! Check the issues for good first tasks.
License
This project is licensed under the MIT License. See the LICENSE.md file for details.