I build native software for Apple platforms—useful beyond the demo, dependable in real work, and sustainable for one developer to maintain.

我为 Apple 平台构建原生软件：不只在演示里成立，也要在真实使用中可靠，并且能够由一个开发者长期维护。

Swift · UIKit · Mac Catalyst · local-first AI

Follow the source. Return to the lecture.

Turn lecture replays into a traceable study path: on-device transcription, exam-focused key points linked back to the teacher's exact words, and LaTeX lecture-note PDFs.

把课程回放整理成一条可追溯的复习路径：在设备上完成转写，把考试重点连回老师原话，并生成 LaTeX 讲义 PDF。

Built as a native Mac Catalyst app with UIKit and whisper.cpp. Optional AI analysis uses an endpoint configured by the user.

使用 UIKit 与 whisper.cpp 构建的原生 Mac Catalyst App；可选的 AI 分析使用用户自主配置的接口。

Repository · Latest release · Feedback

• Native by default · 原生优先 — Respect the platform, its interaction patterns, and the details users feel every day. / 尊重平台、交互习惯，以及用户每天都会感受到的细节。
• Local where it matters · 该在本地，就留在本地 — Prefer on-device work for privacy, resilience, and predictable operating costs. / 为了隐私、可靠性与可控成本，尽可能在设备上完成核心工作。
• Quality before reach · 质量先于流量 — AI can accelerate the work; it does not replace product judgment or engineering responsibility. / AI 可以加速开发，但不会代替产品判断与工程责任。
• Clear promises · 清晰承诺 — Build focused products with understandable boundaries and maintenance expectations. / 做边界明确的产品，也明确说明维护预期。

UIKit is my main tool today. I am bringing SwiftUI into the toolbox while continuing to study the systems and engineering beneath the interface.

UIKit 是我目前的主力工具。我也在逐步把 SwiftUI 纳入工具箱，并持续理解界面之下的系统与工程原理。

Feedback grounded in real use is always welcome. / 始终欢迎来自真实使用的反馈。

A heart that longs to set off needs no justification.

一颗想要出发的心，不需要理由。