I build tools where audio stays on your device and APIs stay under your control — offline voice input on macOS & Android, and self-hosted developer proxies in Rust.
做端侧语音识别和本地优先的开发工具:语音不上云,接口可自托管。
| Project | What it does | |
|---|---|---|
| 🎤 | LocalVoice | Offline macOS dictation — push-to-talk, MLX + Qwen3-ASR, bilingual CN/EN, optional AI rewrite |
| 📱 | Fcitx5 SenseVoice | Offline Chinese voice IME for Android — sherpa-onnx, Silero VAD, SenseVoice, no network permission |
| 🦀 | kiro.rs | Rust proxy: Anthropic Messages & OpenAI-compatible APIs → Kiro / Amazon Q, with Web Admin |
- Blog (中文): stormrise.github.io
- Open to issues & PRs on the projects above
Building speech tools that don't phone home.

