I put AI inference on devices, and terminals on Vision Pro.
- llama.rn
β React Native binding for llama.cpp. LLMs running fully on-device, no server.
- llama.node β llama.cpp for Node.js / WASM, same llama.rn API.
- whisper.rn
β React Native binding for whisper.cpp. On-device speech recognition.
- whisper.node β whisper.cpp for Node.js / WASM, same whisper.rn API.
- Collaborator on llama.cpp (31 PRs merged) and contributor to whisper.cpp (14) β most of that work flows straight back into the bindings above.
- Multiplex β an SSH / tmux / herdr (coming soon) terminal built for Apple Vision Pro, iPad and iPhone. On the App Store, going Apache-2.0 soon.
Building BRICKS β an application system for interactive content β and CTOR, our agent harness & desktop IDE for it.
react-native-debugger (β 10k, unmaintained) Β· redux-devtools Β· electron-react-boilerplate
ποΈ Big 3 and old-school training Β· π Japanese (JLPT N2+) Β· π¦ @jhen_jie