SuperVoiceMode dictation experiment became an AI voice interface
Local transcription with AI correction beats cloud latency, but voice assistants are crowded.

1.2% WER in 150ms beats Whisper and Deepgram, but pricing undercuts adoption vs free.
Developers, writers, accessibility users seeking fast, private speech-to-text
Whisper · Deepgram · AssemblyAI
Local transcription with AI correction beats cloud latency, but voice assistants are crowded.
One Rust binary does what Electron apps and Python scripts couldn't for Linux dictation.
Nine personality modes are prompt variations wrapped in Tkinter with Groq API.
Fully-local Whisper+Ollama voice stack; real Windows app with emoji widget.
Sub-sentence TTS streaming beats Piper/Sherpa-ONNX latency by token-level triggering on CPU.
Linux finally gets offline voice typing; Ctrl-tap + Vulkan GPU support vs cloud-dependent alternatives.