vocalinux and voxtype
These are complements: vocalinux provides multiple speech recognition backends (Whisper.cpp, Whisper, VOSK) that could power voxtype's push-to-talk interface, which lacks its own specified recognition engine.
About vocalinux
jatinkrmalik/vocalinux
Free, open-source, 100% offline voice dictation for Linux. Speak and type anywhere via whisper.cpp, Whisper & VOSK engines, GPU-accelerated, works on X11 + Wayland!
Provides intelligent hardware detection and multi-engine flexibility: users can choose between whisper.cpp (default, universal Vulkan GPU support), OpenAI Whisper (PyTorch-based, NVIDIA-only), or VOSK (lightweight). Integrates with X11/Wayland input systems via IBus and evdev for keyboard event injection, system tray via AppIndicator, and XDG autostart for persistent startup. Architecture emphasizes reliability with SYN_DROPPED recovery, sample-rate device probing, and modifier alias matching across keyboard layouts.
About voxtype
peteonrails/voxtype
Voice-to-text with push-to-talk for Wayland compositors
Supports 7 offline transcription engines (Whisper, Moonshine, SenseVoice, etc.) with native CJK language support, meeting mode for continuous transcription with speaker attribution and multi-format export, and flexible text input via compositor keybindings or evdev fallback. Architecture uses whisper.cpp for local inference with configurable post-processing pipelines, optional Ollama integration for grammar correction, and multiple typing backends (wtype, dotool, ydotool, clipboard) with automatic fallback chains based on environment capabilities.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work