Turn speech into text anywhere via hotkey (runs on Intel NPU, no cloud)
Intel NPU offload frees GPU/CPU, but Windows-only with tiny hardware audience.
Speech to text at cursor using NPU. (shortcut -> win+/)
Offloads Whisper to Intel NPU so GPU stays free for other tasks.
Windows users with Intel NPU hardware (Core Ultra processors)
Superwhisper · Whisper.cpp · MacWhisper
Intel NPU offload frees GPU/CPU, but Windows-only with tiny hardware audience.
$39 wrapper around Whisper API when Superwhisper and Wispr Flow already dominate this space.
Zero-knowledge encryption for transcripts is nice, but Otter.ai and TurboScribe exist.
The repo actually solves the messy plumbing of live voice agents: modular ASR→LLM→TTS adapters plus an optional PersonaPlex speech-to-speech path, per-agent env overrides, and a Playwright-driven Jitsi bot for room joining. It's a useful MVP for anyone prototyping AI co-hosts, though mixing backends is still manual and PersonaPlex demands extra infra, so it's more pragmatic experiment than turnkey product.
Brings your own model and stays out of the way, unlike bloated SaaS alternatives.
Local Whisper with auto-paste and hotkey—Cursor/Claude-coded, shipped, and works.